Dopp RDR WX Obs 06

U.S.
DEPARTMENT OF COMMERCE/ National Oceanic and Atmospheric Administration
OFFICE OF THE FEDERAL COORDINATOR FOR

METEOROLOGICAL SERVICES AND SUPPORTING RESEARCH
FEDERAL METEOROLOGICAL HANDBOOK NO. 11
DOPPLER RADAR
METEOROLOGICAL
OBSERVATIONS
PART D
WSR-88D UNIT DESCRIPTION
AND OPERATIONAL
APPLICATIONS
FCM-H11D-2006
Washington, DC
February 2006
THE FEDERAL COMMITTEE FOR
METEOROLOGICAL SERVICES AND SUPPORTING RESEARCH (FCMSSR)
VADM CONRAD C. LAUTENBACHER, JR., USN (RET.) MR. RANDOLPH LYON

Chairman, Department of Commerce Office of Management and Budget
DR. SHARON HAYS (Acting) MR. CHARLES E. KEEGAN

Office of Science and Technology Policy Department of Transportation
DR. RAYMOND MOTHA MR. DAVID MAURSTAD (Acting)

Department of Agriculture Federal Emergency Management Agency
Department of Homeland Security
BRIG GEN DAVID L. JOHNSON, USAF (RET.)
Department of Commerce DR. MARY L. CLEAVE
National Aeronautics and Space
MR. ALAN SHAFFER Administration
Department of Defense
DR. MARGARET S. LEINEN
DR. ARISTIDES PATRINOS National Science Foundation
Department of Energy
MR. PAUL MISENCIK
DR. MAUREEN MCCARTHY National Transportation Safety Board
Science and Technology Directorate
Department of Homeland Security MR. JAMES WIGGINS
U.S. Nuclear Regulatory Commission
DR. MICHAEL SOUKUP
Department of the Interior DR. LAWRENCE REITER
Environmental Protection Agency
MR. RALPH BRAIBANTI
Department of State MR. SAMUEL P. WILLIAMSON
Federal Coordinator
MR. JAMES B. HARRISON, Executive Secretary

Office of the Federal Coordinator for
Meteorological Services and Supporting Research
THE INTERDEPARTMENTAL COMMITTEE FOR

METEOROLOGICAL SERVICES AND SUPPORTING RESEARCH (ICMSSR)
MR. SAMUEL P. WILLIAMSON, Chairman MS. LISA BEE

Federal Coordinator Federal Aviation Administration
Department of Transportation
MR. THOMAS PUTERBAUGH
Department of Agriculture DR. JONATHAN M. BERKSON
United States Coast Guard
MR. JOHN E. JONES, JR. Department of Homeland Security
Department of Commerce
MR. JEFFREY MACLURE
RADM FRED BYUS, USN Department of State
United States Navy
Department of Defense DR. S. T. RAO
Environmental Protection Agency
COL JOHN D. MURPHY, USAF (Acting)
United States Air Force MR. JOHN GAMBEL
Department of Defense Federal Emergency Management Agency
Department of Homeland Security
MR. RICKEY PETTY
Department of Energy DR. RAMESH KAKAR
National Aeronautics and Space
MR. CHRISTOPHER DOYLE Administration
Science and Technology Directorate
Department of Homeland Security DR. JARVIS MOYERS
National Science Foundation
MR. JOHN VIMONT
Department of the Interior MR. DONALD E. EICK
National Transportation Safety Board
MS. REGINA MCELROY
Federal Highway Administration MS. LETA A. BROWN
Department of Transportation U.S. Nuclear Regulatory Commission
MS. ANDREA PETRO

Office of Management and Budget
MR. JAMES B. HARRISON, Executive Secretary

Office of the Federal Coordinator for
Meteorological Services and Supporting Research
FEDERAL COORDINATOR FOR
METEOROLOGICAL SERVICES AND SUPPORTING RESEARCH
DOPPLER RADAR
METEOROLOGICAL OBSERVATIONS
FEDERAL
METEOROLOGICAL
HANDBOOK NO. 11
PART D
WSR-88D UNIT DESCRIPTION
AND OPERATIONAL APPLICATIONS
FCM-H11D-2006
Washington, DC
FEBRUARY 2006 FMH-11-PART D

PREFACE
The Federal Coordinator for Meteorological Services and Supporting Research has the responsibility to
maintain and publish Federal Meteorological Handbooks. This series of documents provides standards
and procedures to facilitate the efficient collection, sharing, and use of meteorological information by
agencies of the federal government and private industry.
The original Federal Meteorological Handbook, Number 11 (FMH-11), DOPPLER RADAR

METEOROLOGICAL OBSERVATIONS, was prepared and published under the auspices of the
Office of the Federal Coordinator for Meteorological Services and Supporting Research (OFCM) at the
request of the Next-Generation Weather Radar (NEXRAD) Program Council and in coordination with
the federal agencies that are represented on the Interdepartmental Committee for Meteorological
Services and Supporting Research. The purpose of FMH-11 is to standardize, insofar as practical, the
operation of the Weather Surveillance Radar-1988, Doppler (WSR-88D) systems and the procedures
used by personnel of the Departments of Commerce, Defense, and Transportation. By approving
publication of this handbook, those agencies have agreed to operate their WSR-88D systems
accordingly. Some flexibility under certain meteorological, siting, or mission circumstances is
permitted to enhance the quality and utility of some WSR-88D products.
The revision process is dependent on the evolution of WSR-88D subsystems software and products.
Part A has been revised in conjunction with recent WSR-88D software releases to ensure it provides
users current operations guidance. Parts B, C, and D have/are being revised in a separate effort
principally through the guidance of the Radar Operations Center (ROC). All revisions are coordinated
among the NEXRAD triagencies (Department of Commerce (DOC), Department of Defense (DoD),
and Department of Transportation (DOT)); thus, they possess the same authority as the initial edition
of FMH-11.
Agencies should review the documents annually. The goal is to review and update the handbooks (as
necessary) as part of every WSR-88D software build release. Suggestions for modifications and
additions should be sent to the WSR-88D Radar Operations Center webmaster at
http://www.roc.noaa.gov/Feedback/ for consideration for inclusion in a future update. As required, the
handbooks will be updated and published in electronic format, and made available on the OFCM home
page at http://www.ofcm.gov/homepage/text/pubs.htm. Readers can make copies of the handbooks
without a request for approval from the OFCM. A summary of changes made during updates will be
annotated in the preface of each part.
Each major part of the FMH-11 is designed to stand alone; however, cross references are used to avoid
unnecessary redundancy. In all, FMH-11 has four parts:
Part A - System Concepts, Responsibilities, and Procedures (December 2005)
Part B - Doppler Radar Theory and Meteorology (December 2005)
Part C - WSR-88D Products and Algorithms (Expected in 2006)
Part D - WSR-88D Unit Description and Operational Applications (February

2006)

ii
Part D describes the main components of the WSR-88D, outlines the flow of data through the unit, and
addresses the main objective of providing Doppler radar meteorological observations. The information
presented provides the user with essential knowledge to enable him or her to better understand and
manage the WSR-88D data resources. Part D provides operational guidance on the use of WSR-88D
products to enhance the observance of meteorological or hydrological phenomena.
Summary of Changes:
This version of Part D replaces the original document, published in April 1992. This version
incorporates the many system changes made since April 1992, updates the document as of Radar
Product Generator (RPG) Build 6 (released in September 2004), and adds information based on the
operational experience gained during the first decade of WSR-88D operational use and case studies.
Samuel P. Williamson
Federal Coordinator for Meteorological
Services and Supporting Research

iii
FEDERAL METEOROLOGICAL HANDBOOK NO. 11
DOPPLER RADAR METEOROLOGICAL OBSERVATIONS
PART D
WSR-88D UNIT DESCRIPTION AND OPERTIONAL APPLICATIONS
TABLE OF CONTENTS
Page
PREFACE ii
TABLE OF CONTENTS iv
LIST OF FIGURES ix
LIST OF TABLES xi
CHAPTER 1. INTRODUCTION
1.1 Background 1-1
1.2 Purpose and Scope 1-1
1.3 Organization 1-1
CHAPTER 2. UNIT DESCRIPTION

2.1 Components 2-1
2.2 Radar Data Acquisition 2-1
2.2.1 Antenna 2-1

2.2.2 Pedestal 2-3
2.2.3 Radome 2-3
2.2.4 Facilities 2-3
2.2.5 Transmitter and Receiver 2-3
2.2.6 Signal Processor 2-4
2.2.7 User Interfaces 2-7
2.2.8 Communications 2-8
2.3 Radar Product Generator 2-8
2.3.1 Master System Control Function 2-8

2.3.2 Archive II and III 2-12
2.3.3 Radar Product Generator Communications Equipment 2-12
2.4 Associated Principal User Display Systems 2-13
2.4.1 Advanced Weather Interactive Processing System 2-13

2.4.2 Weather and Radar Processor 2-13
2.4.3 Integrated Terminal Weather System 2-13
2.4.4 Open System Principal User Processor 2-13
FEBRUARY 2006 FMH-11-PART-D

iv
Page
2.5 Communications 2-14
2.5.1 Wideband Communications 2-14

2.5.2 Narrowband Communications 2-15
CHAPTER 3. UNIT FUNCTIONAL FLOW

3.1 Introduction 3-1
3.2 Radar Data Acquisition 3-1
3.2.1 Handle Maintenance Console 3-1

3.2.2 Monitor and Calibrate Radar Data Acquisition 3-3
3.2.3 Form Radar Data Acquisition Data 3-4
3.2.4 Handle Wideband Data 3-4
3.2.5 Control Radar Data Acquisition 3-5
3.2.6 Control Signal Processor 3-6
3.2.7 Control Pedestal 3-7
3.3 Radar Product Generation 3-8
3.3.1 System Control 3-10

3.3.2 Base Data Acquisition 3-10
3.3.3 Preprocess Base Data 3-11
3.3.4 Product Generation 3-15
3.3.5 Product Distribution 3-26
CHAPTER 4. OPERATIONAL APPLICATIONS

4.1 Introduction 4-1
4.2 Non-Meteorological Radar Echoes 4-1
4.2.1 Ground Clutter 4-1

4.2.2 Ground Clutter Returns from Anomalous Propagation 4-3
4.2.3 Sidelobes 4-5
4.2.4 Solar Effects 4-6
4.2.5 Range Folded Data 4-8
4.2.6 Velocity Aliased Data 4-11
4.3 Non-Precipitation Radar Echoes 4-13
4.3.1 Recognition of Non-Precipitation Radar Echoes 4-15

4.3.2 Considerations 4-15
4.4 Boundaries 4-15
4.4.1 Recognition of Boundaries 4-15


v
Page
4.5 Cloud Layers 4-19
4.5.1 Recognition of Cloud Layers 4-20

4.6 Pre-Convective Development 4-22
4.6.1 Recognition of Pre-Convective Development 4-22

4.7 Convective Storms 4-22
4.7.1 Single Cell Storms 4-22

4.7.2 Multicell Storms 4-27
4.7.3 Supercell Storms 4-31
4.7.4 Severe Thunderstorm Identification 4-42
4.8 Hail 4-65
4.8.1 Recognition of Hail Potential 4-65

4.9 Extratropical Cyclones 4-70
4.9.1 Evolution of Stratiform Clouds and Precipitation 4-71

4.9.2 System Movement 4-74
4.9.3 Cell Movement 4-75
4.10 Winter Storms 4-75
4.10.1 Onset of Surface Precipitation 4-75

4.10.2 Cold Air Depth, Freezing Level 4-77
4.10.3 Snowfall Estimation 4-80
4.10.4 Recognition of Heavy Snow and Snow Showers 4-84
4.10.5 Rain/Snow Line 4-89
4.11 Tropical Cyclones 4-89
4.11.1 Embedded Convective Storms and Tornadoes 4-91

4.11.2 Heavy Rainfall 4-92
4.11.3 Convective Scale and Straight Line Winds 4-95

vi
Page
4.12 Flash Floods 4-99
4.12.1 Recognition of a Flash Flood Event 4-99

4.12.2 Precipitation Accumulation 4-102
4.13 Non-Convective High Wind Events 4-106
4.13.1 Recognition of High Wind Events 4-106

4.14 Low-Level Jets 4-107
4.14.1 Recognition of Low-Level Jets 4-107

4.15 Wind Shear 4-110
4.15.1 Recognition of Wind Shear 4-110

4.16 Turbulence 4-110
4.16.1 Recognition of Turbulence 4-113

4.17 Microbursts 4-113
4.17.1 Recognition of Microbursts 4-113

4.18 Considerations for Building a Routine Product Set 4-118
4.18.1 Initial Considerations 4-118

4.18.2 Some Local RPS List Considerations 4-118
4.18.3 Product Considerations 4-119
4.18.4 National RPS List Requirements 4-120
4.18.5 Clear Air Mode 4-121
4.18.6 Precipitation Mode 4-121

vii
Page
4.19 Automated Alerts 4-122
4.19.1 Alert Use 4-122

4.19.2 Setting Alerts 4-123
4.19.3 Alert Groups 4-124
4.19.4 Alert Areas and Acknowledgments 4-124
4.19.5 User Alert Message (UAM) Product 4-125
4.19.6 Alert-Paired Products 4-125
APPENDIX A. ACRONYMS AND ABBREVIATIONS A-1
APPENDIX B. GLOSSARY B-1

viii
LIST OF FIGURES
Figure Page
2-1. RPG Graphical User Interface 2-10
3-1. Radar Data Acquisition Status and Control Program Data Flow 3-2
3-2. Radar Product Generation Data Flow 3-9
3-3. Types of Load Shedding at the Radar Product Generator 3-12
3-4a. Flow of Data for Generation of Products 3-18
3-4b. Flow of Data for Generation of Products (Continued) 3-19
3-4c. Flow of Data for Generation of Products (Continued) 3-20
3-4d. Flow of Data for Generation of Products (Continued) 3-21
3-4e. Flow of Data for Generation of Products (Continued) 3-22
3-4f. Flow of Data for Generation of Products (Continued) 3-23
3-4g. Flow of Data for Generation of Products (Concluded) 3-24
4-1. Residual Ground Clutter 4-2

4-2. Anomalous Propagation 4-4
4-3. Solar Effects 4-7
4-4. Range Overlaid Echoes 4-9
4-5. Velocity Dealiasing Errors 4-12
4-6. Dust Storm Reflectivity 4-14
4-7. Dryline Reflectivity 4-16
4-8. Boundaries and Severe Convection Reflectivities 4-17
4-9. Dryline – Cold Front Intersection Reflectivity 4-18
4-10. Clear Air Reflectivity Product 4-21
4-11. Ordinary Cell 4-24
4-12. Non-Severe Thunderstorm Schematic 4-25
4-13. Ordinary Non-Severe Thunderstorm 4-26
4-14. Multicell Storm Sequence Schematic 4-28
4-15. Severe Thunderstorm Schematic 4-30
4-16. Left-Moving Severe Thunderstorm Reflectivity 4-34
4-17. Classic Supercell Thunderstorm Schematic 4-35
4-18. Supercell Thunderstorm Reflectivity 4-36
4-19. High Precipitation Supercell Thunderstorm 4-38
4-20. Reflectivity and Mean Radial Velocity Display of a High
Precipitation Supercell 4-39
4-21. Mini-Supercell Thunderstorm Reflectivity 4-43
4-22. Low Precipitation Supercell Thunderstorm Reflectivity 4-44
4-23. Mature Severe Thunderstorm Schematic 4-47
4-24. Tornadic Supercell Thunderstorm Reflectivity 4-49
4-25. Bow Echo Schematic 4-51
4-26. Squall Line Schematic 4-52
4-27. Mesocyclones 4-56
4-28. Combined Rankine Vortex Schematic 4-57
4-29. Mesocyclone and TVS Vertical Extent 4-58

ix
Figure Page
4-30. Multiple-Core Mesocyclonic Storm Schematic 4-60

4-31. Tornado Vortex Signature Vertical Extent 4-62
4-32. Storm Summit Divergence vs. Hailstone Size 4-67
4-33. Radar Signal Path Schematic 4-68
4-34. Three-Body Scatter Spike 4-69
4-35. National Mosaic of Base Reflectivity of an Extratropical Cyclone 4-72
4-36. Regional Mosaic of Base Reflectivity of an Extratropical Cyclone 4-73
4-37. Precipitation Mode Reflectivity Scan of Virga 4-76
4-38. VAD Wind Profile Product 4-78
4-39. Bright Band 4-79
4-40. One-Hour Snow Depth Accumulation Product 4-81
4-41. One-Hour Snow Water Equivalent Accumulation Product 4-82
4-42. Storm Total Snow Depth Accumulation Product 4-83
4-43. User Selectable Snow Accumulation Product 4-85
4-44. User Selectable Snow Water Equivalent Accumulation Product 4-86
4-45. Convective Snowfall Reflectivity 4-87
4-46. Hurricane Charley Reflectivity 4-90
4-47. Mini-Supercell and Mesocyclones Associated with Hurricane Ivan 4-93
4-48. Storm Total Rainfall Accumulation Product of Hurricane Emily 4-94
4-49. Mean Radial Velocity Product of Hurricane Charley 4-96
4-50. WSR-57 Radar Reflectivity of Hurricane Andrew 4-97
4-51. Surface Observations of Hurricane Andrew 4-98
4-52. One-Hour Rainfall Accumulation Product 4-100
4-53. Storm Total Rainfall Accumulation Product 4-101
4-54. Three-Hour Rainfall Accumulation Product 4-103
4-55. User Selectable Rainfall Accumulation Product 4-104
4-56. Smoke Plume Detected by Radar 4-108
4-57. Base Velocity Data Array Product of a Low-Level Jet 4-109
4-58. Severe Weather Analysis (Velocity) Product 4-111
4-59. Severe Weather Analysis (Spectrum Width) Product 4-112
4-60. Microburst Detected on a Mean Radial Velocity Product 4-115
4-61. Reflectivity Cross Section of a Microburst-Producing Thunderstorm 4-116
4-62. Velocity Cross Section of a Microburst-Producing Thunderstorm 4-117

x
LIST OF TABLES
Table Page
2-1. Relationships 2-2
4-1. Alert Types and Mnemonics (With the Assigned Order of Importance) 4-126
4-2. Alert Locations and Paired Window Product Centers 4-127
4-3. Application versus Product 4-128

xi
CHAPTER 1
INTRODUCTION
1.1 Background. The material presented in Part D, WSR-88D Unit Description and Operational
Applications, of Federal Meteorological Handbook No. 11 (FMH-11) is based on data and
information gathered from several sources. Unit description and functional flow material were
extracted from technical manuals and computer program specifications prepared by the original
system contractor and updated by the Radar Operations Center (ROC) and NEXRAD agencies.
The material in Part D is as of software Build 6 for the Radar Product Generator (RPG) and Open
Principal User Processor (OPUP), Build 10.2 of the legacy Radar Data Acquisition (RDA), and
Build 10 of the legacy Principal User Processor (PUP). Chapter 4, dealing with operational
applications, is based primarily on the inputs of agency personnel familiar with the operation and
meteorological use of the WSR-88D, the National Weather Service’s (NWS) Warning Decision
Training Branch (WDTB), ROC subject matter experts, and a support services contractor.
1.2 Purpose and Scope. Part D is intended to provide the WSR-88D user community with an
overview of unit operations and functional flow and insights into the meteorological applications of
the WSR-88D. Additional and more detailed information regarding operating instructions are
contained in baseline WSR-88D technical manuals. Recommended changes/corrections to this
manuscript are welcome and should be sent to: http://www.roc.noaa.gov/Feedback/.
1.3 Organization. Part D covers three main topics of unit operations. They are:
Chapter 2, Unit Description
Chapter 3, Unit Functional Flow
Chapter 4, Operational Applications
The listings of site-adaptable parameters for the RDA, RPG, and PUP that were in Appendices A,
B, and C of the April 1992 version of Part D can now be found in the Guidance on Adaptable
Parameters at: http://www.roc.noaa.gov/ssb/sysdoc/Operations.asp. These documents are updated
with each WSR-88D software release. Part A of this Handbook defines the authority to make
changes to these parameters.
The WSR-88D images used in Part D are from the many diverse NEXRAD agency user display
systems available. The list of the primary user display systems is in Section 2.4. In addition,
figures from the proof-of-concept National Severe Storms Laboratory (NSSL) Warning Decision

1-1
Support System (WDSS), the National Climatic Data Center (NCDC) Level III/product archives
via the NCDC NEXRAD Viewer, and the Common Operations and Development Environment
(CODE) CODEview Graphics are used. These diverse displays are used to ensure the best
representative examples of the phenomena being depicted are presented.
Appendix A lists acronyms and abbreviations found in Part D. Appendix B is a glossary of terms
used in Part D.

1-2
CHAPTER 2
UNIT DESCRIPTION
2.1 Components. This chapter contains a description of the major components of a WSR-88D
unit. The major components are:
Radar Data Acquisition

Radar Product Generator
Associated Principal User Display Systems
Communications
Each component is discussed in general terms in this chapter. For a more detailed description,
refer to the WSR-88D technical manuals. In the case of associated principal user display systems,
refer to the technical manuals of the respective display system.
2.2 Radar Data Acquisition. The RDA provides the detection and measurement of
meteorological and hydrological phenomena. The RDA functional area consists of the hardware,
firmware, and software required to control the antenna, transmit and receive radio frequency
pulses, perform signal processing, ground clutter suppression, range unfolding and distribute base
data in the form of three base moments (reflectivity, mean radial velocity, and velocity spectrum
width) to the RPG. The RDA also provides the capability for calibration, status monitoring, and
error detection in support of both operations and maintenance activities (both local and through the
Federal Aviation Administration (FAA) Remote Monitoring Subsystem (RMS)).
The RDA is composed of the antenna, pedestal, radome, facilities, transmitter, receiver, signal
processor, user interfaces, and communications. Basic information on each of these is provided
below. Additional detail can be found in the RDA technical manuals and Part B of this Handbook.
2.2.1 Antenna. The antenna uses a parabolic reflector, 8.5 m (28 ft) in diameter, to
provide a beam width of approximately 0.93o. The antenna feed horn applies linear horizontal
polarization to transmit and receive S-band radio frequency (rf) pulses. Return signals from
weather echoes are routed to the receiver and converted for use by the signal processor.
2.2.1.1 Polarization. In addition to frequency, an electromagnetic wave has a

property called its state of polarization, which describes the behavior of the electric field
component. The electric field component of an electromagnetic wave is perpendicular to the
direction of wave propagation. The constantly changing orientation of the electric field component
with respect to the direction of propagation defines the polarization of the electromagnetic wave.
In general, the electric field traces out an ellipse as the wave propagates through space. Circular
and linear polarizations are special cases of polarization of interest in radar technology. Circular
polarization is characterized by an electric field of constant magnitude that rotates around the axis
of propagation; linear polarization is characterized by an electric field that oscillates in a plane.
FEBRUARY 2006 FHM-11-PART-D

2-1
Linear polarization can be further characterized by the orientation of the plane of oscillation, e.g.,
linear horizontal or linear vertical. An electromagnetic wave is said to be unpolarized when the
electric field component appears to fluctuate randomly.
Polarized waves are used in radar applications to improve target detection. The WSR-88D
employs linear horizontal polarization to enhance liquid precipitation detection. As liquid
precipitation falls, it tends to flatten, giving it the appearance of an oblate spheroid. Linear
horizontal polarization takes advantage of the resultant exaggerated horizontal cross-section of the
rain drop.
2.2.1.2 Volume Coverage Patterns. To adequately sample the atmosphere within

the range of the radar, the WSR-88D unit employs several scanning strategies, which are
implemented as Volume Coverage Patterns (VCPs). A VCP is a series of elevation angles
completed within a specified period of time. Each VCP has distinct elevations, Pulse Repetition
Frequency (PRF), and rotational speed requirements. The VCPs designed for precipitation
detection implement three distinct waveforms to allow for long-range detection of reflectivity and a
high Nyquist interval to reduce velocity aliasing. In general, the lowest two elevation angles are
scanned twice. The first scan uses a contiguous surveillance waveform and the second uses a
contiguous Doppler waveform. This is done to resolve ambiguities of velocity aliasing and range
folding and to allow for better clutter suppression. The middle elevation scans employ a combined
waveform called "batch mode" that provides estimates of reflectivity and Doppler moments. The
higher elevation scans use a contiguous Doppler waveform without range unfolding since
measurements of reflectivity at long range are not required at those elevations. (See Part C,
Chapter 5, of this Handbook for a more complete explanation and detailed description of these
scans.)
Table 2-1 shows the relationships between long and short pulses, Precipitation and Clear Air
Modes, and surveillance and Doppler data sampling.
Table 2–1
Relationships
Data Averaging (Precipitation Mode) [samples per bin]

Surveillance 3-28
Doppler 25-111
Data Averaging (Clear Air Mode) [samples per bin]

Surveillance (long pulse) 63
Doppler (long pulse) 87
Surveillance (short pulse) 11-64
Doppler (short pulse) 220-278
2.2.1.2.1 Short Pulse. Short pulse is used in VCP 32 in the Clear Air Mode
(scanning of 5 elevation angles completed in 10 minutes); and VCPs 11 (14 elevation angles in 5
minutes), 12 (14 elevation angles in 4.1 minutes), 21 (9 elevation angles in 6 minutes), and 121 (9

2-2
elevation angles in 5 minutes) in the Precipitation Mode. The pulse length is 1.57 µs or ~250 m
and is used in conjunction with PRFs ranging from 318 to 1304 pulses per second.
2.2.1.2.2 Long Pulse. Long pulse is used in VCP 31 when operating in the Clear
Air Mode. As in VCP 32, 5 elevation angles are scanned in 10 minutes. The long pulse is 4.7 µs
or ~0.75 km (0.4 nm) and is used in conjunction with a PRF of 318 to 452 pulses per second. Long
pulse results in enhanced sensitivity and can measure weaker return signals. Using the simple ratio
of the increase in pulse depth as a measure of the increased energy provided to the return, we could
expect an increase in signal to noise ratio of about 4.2 dB. In addition, the system attempts to
provide an equivalent to a matched filter for the long pulse through an averaging process.
However, due to losses in the process rather than achieving the theoretical increase of 8.4 dB, the
increase in sensitivity of long pulse to short pulse is approximately 6 dB.
2.2.2 Pedestal. The antenna pedestal is an aluminum and cast iron structure that positions
the parabolic antenna in both azimuth and elevation. In an operational mode, the antenna rotates
continuously in azimuth at a maximum speed of 5 revolutions per minute (RPM) (while the
pedestal is capable of 6 rpm) and moves in incremental steps in elevation at the rate of one step per
azimuth revolution. The physical limits on the antenna elevation are -1o and +60o. The operational
VCPs cover elevation angles from 0.5o to 19.5o.
2.2.3 Radome. The radome, which protects the antenna and pedestal from weather
hazards, is of rigid fiber glass construction with a diameter of 11.9 m (39 ft) and an rf two-way loss
of 0.6 decibel (dB) at 2,800 megaHertz (MHz). Its design inhibits ice and snow buildup. The
radome is coated with a hydrophobic paint which sheds water readily. The radome has
maintenance heating, forced air ventilation, and aircraft warning lights, when required. The
radome is also equipped with a lightning protection system consisting of either a single air terminal
and associated grounding system or an array of multiple air terminals for selected configurations.
2.2.4 Facilities. The facilities include buildings, towers, power generation, and air
conditioning equipment. The WSR-88D tower is of steel construction and assembled in increments
of 5 m. The engine generator is housed in a separate enclosure, as is the RDA and RPG (where
collocated), and any redundant radar and support equipment. Support equipment for the tower
includes prime power distribution, backup power, a transition power system, environmental control
(heating, ventilating, and air conditioning), and remote sensors for security and environmental
equipment failures.
2.2.5 Transmitter and Receiver. The transmitter is a high-power, S-band, pulse amplifier
that generates the rf pulse for transmission via the antenna. Input to the transmitter is a gated, low-
power (10 milliwatts) rf drive signal with a frequency range of 2,700 to 3,000 MHz, which is
generated by the receiver. The signal provides coherency (phase reference) for the return signal
from which to compute the Doppler phase shift. The Klystron transmitter then transmits a burst of
high-power rf energy of either 1.57 or 4.7 microseconds (µs) in duration, with a peak operating
power of approximately 750 kilowatts. The high-power rf energy is transmitted to the antenna
through a waveguide.

2-3
The receiver sends the rf amplified signal and synchronization signals to the transmitter, amplifies
the rf signal from the antenna, and transfers the analog data to the signal processor. It sends signals
to and receives commands from the RDA Status and Control (RDASC) program.
2.2.6 Signal Processor. The signal processor accepts the analog data from the receiver
and processes the data into base data that are sent to the RDA control. The processing is
accomplished using a hardwired signal processor (HSP) and a programmable signal processor
(PSP).
2.2.6.1 Hardwired Signal Processor. The functions of the HSP are to convert
analog data into digital data, perform clutter filtering, and maintain system synchronization. The
HSP processes and outputs the radar data on a 0.25 km (0.13 nm) gate-by-gate basis and sends
these data to the PSP.
2.2.6.1.1 Analog to Digital Conversion. The analog-to-digital (A/D) converter

assembly is physically part of the receiver (for low noise considerations), but is functionally
considered to be part of the signal processor. The A/D converter accepts analog in-phase (I) and
quadrature (Q) and logarithmic (LOG) inputs and converts them to 12 bit two's complement at a
0.6 MHz rate (0.25 km (0.13 nm) sampling interval). The converter has test/bias digital inputs that
provide two functions. During normal operations, these inputs are used to null out direct current
biases, which would degrade base data. During A/D tests, these inputs inject test signals.
The A/D converter outputs I, Q and LOG on three serial interfaces at a 9.6 MHz rate (burst of 12
bits per range bin). All signal processor interfaces are isolated from the receiver with opto-
isolators to reduce the effects of digital noise on the receiver sensitivity.
2.2.6.1.2 Clutter Filtering. The WSR-88D provides for the suppression

(reduction) of returned power whose radial velocity falls within the specified "notch width" (range
of velocities centered on zero) within predefined areas. Basically, the larger the notch width the
greater is the reduction in near-zero velocity signal power (suppression). This reduction in signal
power effectively decreases the clutter's power contribution in the given range bin. To maintain
meteorological return integrity, only the signal power whose radial motion falls within the notch
width is reduced.
The HSP uses a site-unique Bypass Map (built by the RDA System Operability Test (RDASOT)
software), Default Notch Width Map (Master System Control Function (MSCF)-controlled RDA
adaptation data), and operator-defined Clutter Suppression Regions to determine the areas in which
to invoke suppression, and the amount of suppression (signal reduction) to apply.
The Bypass Map and Default Notch Width Map are used to identify and suppress areas of known
ground clutter return. In the absence of any operator-defined Clutter Suppression Regions, the
WSR-88D will use the Bypass Map to determine where to apply clutter suppression and the
Default Notch Width Map notch width definitions (stored RDA adaptation data) to determine the
amount of suppression to be applied.

2-4
In accordance with Unit Radar Committee (URC) guidelines, as specified in Part A of this
Handbook, the MSCF operator may modify the application of clutter suppression to known ground
clutter areas and transient areas through the definition of Clutter Suppression Regions.
Inappropriate application of operator-defined Clutter Suppression Regions can result in degradation
of data accuracy and quality.
The WSR-88D detects only the radial component of the actual target motion. Therefore, if the
target is moving perpendicular to the radar beam it has a zero (or near zero) radial motion. If
clutter suppression has been invoked for a given range bin whose radial target motion is zero (or
near zero), the defined power reduction factor will be applied to that bin. This power reduction
will result in a significantly reduced reflectivity estimate for the range bins whose mean radial
velocity is at or near zero.
In areas with a strong azimuthal reflectivity gradient, clutter suppression may result in an artificial
displacement of the returned power maxima into azimuthally adjacent range resolution cells. This
data smearing is an artifact of the clutter filter response times and occurs when high reflectivity
gradients are present. The clutter filter decay rate is slower with narrower notch widths; therefore,
data smearing is more likely to be seen when low suppression levels are used. When data smearing
occurs, weak reflectivity features on the trailing edge (azimuthally) of a high reflectivity gradient
may be obscured. Additionally, the apparent aerial extent of high reflectivity will be increased.
The impacts of data smearing on the velocity and spectrum width estimates vary depending upon
the relative power contributions from the range resolution cells for the affected range bin.
Appropriate application of clutter suppression, however, will result in several improvements in

WSR-88D data quality and operations. The most obvious benefit attained as a result of proper
clutter suppression is that ground clutter is removed from the graphical products displayed at the
user display system or workstation. However, several more important data quality improvements
can be realized by the optimum application of clutter suppression.
The most important benefit is the improved accuracy of the WSR-88D products. Since clutter
suppression occurs prior to the calculation of the base data estimates, proper clutter filtering will
result in the base data estimates being more representative of the actual meteorological situation.
Consequently, the more accurate the base data estimates, the more reliable the output from
downstream processing and algorithms and, as a result, the more accurate base and derived
products.
By removing (reducing) the power contribution of ground targets from the base data estimate for
first-trip range bins, the likelihood of assigning valid velocity and spectrum width data to second-
trip range bins is increased.
A large number of velocity dealiasing failures that occur in VCP 31 can be attributed to ground
clutter induced bias in the base velocity estimate. The removal of the ground clutter bias from the
base velocity estimate will result in meteorologically plausible ambiguous velocity estimates.
These ambiguous velocity estimates can then be readily dealiased by the dealiasing algorithm.
However, residual ground clutter will continue to cause velocity dealiasing algorithm failures.

2-5
The Clear-Air Mode deselection logic employed by the WSR-88D compares the area coverage of
precipitation-like return to a specified value (nominal clutter area plus precipitation area threshold).
If the area coverage of precipitation-like return exceeds this value, the RPG computer commands
the RDA computer to switch to Precipitation Mode. The reduction of non-meteorological data
from the aerial coverage computation will result in a more accurate estimate of the actual
precipitation present, thereby reducing the likelihood of prematurely switching to Precipitation
Mode due to ground clutter returns.
2.2.6.1.3 System Synchronization. Synchronization provides computer

controlled range counters and range decoding logic to generate timing triggers, clocks, and gates
for the transmitter, receiver, and the signal processor. It also provides computer controlled radial
sequence counters and decoding logic to control radial timing to generate radial clocks. The range
counter and radial counter logic, in combination, generate contiguous surveillance, contiguous
Doppler, and batch waveforms.
For each waveform type, the radial consists of a specified number of pulses. The PRF, the number
of PRFs, and the azimuth scan rate control the azimuthal extent of the radial. Radials normally
have an azimuthal extent of one degree.
2.2.6.2 Programmable Signal Processor. The PSP receives time-series digital

data from the HSP and develops the echo power, reflectivity, mean radial velocity, and spectrum
width arrays. These arrays are sent to the RDASC processor where base data are formatted and
stored. Within the PSP, point clutter suppression and range unfolding are performed. These base
data are then sent back to the RDASC and header information (time, azimuth, elevation, and
calibration parameters) is attached prior to transmission to the RPG.
2.2.6.2.1 Strong Point Clutter Suppression. In the PSP, strong point clutter
suppression is accomplished by computing a digital estimate of the power sum for each sample
volume. The PSP analyzes the power sum array to detect strong point clutter. The PSP replaces
the power data in sample volumes having abnormally high return power with values calculated
from nearby unaffected sample volumes. These power sum array data are then processed to
compute the base data.
2.2.6.2.2 Range Unfolding. The WSR-88D uses three techniques to resolve range
folding. The first technique, used at the lowest two elevation angles, scans each elevation angle
twice. The first scan uses a contiguous surveillance waveform and the second scan uses a
contiguous Doppler waveform. The contiguous surveillance scan generates an echo power array
with 0.5 dB and 1 km (0.54 nm) resolution for each radial. This reflectivity information is sent to
the RDASC as each radial is collected. The echo power arrays are stored in the RDASC until the
end of the surveillance scan and are then sent back to the PSP, radial by radial, while the PSP
processes the contiguous Doppler scan for the same elevation angle. The PSP compares the radial
information collected during the contiguous Doppler scan to the corresponding radial information
collected during the surveillance scan to correctly correlate areas of Doppler data with areas of
reflectivity data. The PSP combines the surveillance and Doppler data into a single record labeled
as "Doppler data."

2-6
The second technique is called “batch mode.” Batch mode employs interlaced surveillance and
Doppler waveforms to achieve the required range while allowing the radar to simultaneously
collect base reflectivity, mean radial velocity, and spectrum width data. This technique is used at
intermediate elevation angles where multiple trip returns can occur in the Doppler waveforms, but
where clutter suppression is not normally required as at the lower elevation angles. Velocity data
are range unfolded in essentially the same process as described above.
Since very little weather occurs above 21.3 km (70,000 ft), velocity data range folding at higher
elevation angles is not an operational problem. Because the height requirement is obtained within
the unambiguous range of the high PRFs, estimates for reflectivity, mean radial velocity, and
spectrum width are achieved using contiguous Doppler waveforms.
The third technique uses the Multiple PRF Dealiasing Algorithm (MPDA). The MPDA scanning
strategy collects sequential scans at the same antenna elevation angle using different Nyquist
velocities (PRFs). These are then range dealiased, aligned, and processed to produce a final
dealiased velocity field based on the combined scans. The actual unambiguous ranges and
corresponding Nyquist velocities depend on the operating frequency of the radar and may be
expressed as follows:
RaVa = c2/(8f)
where c is the speed of light, f is frequency, Ra is the unambiguous range, and Va is the Nyquist
velocity.
The MPDA requires a new VCP, VCP 121. Volume Coverage Pattern 121 requires three Doppler
scans, besides the surveillance scan, at both 0.5° and 1.5° elevation; two additional Doppler Scans
in addition to the batch mode scans at 2.4° and 3.3° elevation; and one additional Doppler scan at
4.3° elevation. Above 4.3° elevation, VCP 121 is identical to VCP 21. Additional information on
VCP 121 can be found in Part C, Chapter 5, of this Handbook.
2.2.7 User Interfaces. The RDA is controlled by the RDASC program. It monitors and
assesses the performance of the RDA, initiates automatic calibration, performs calibration
calculations, and reports RDA status to the RPG. It formats reflectivity, mean radial velocity, and
spectrum width data input from the signal processor, attaches header data, and initiates data
transfer to the RPG.
The RDASC program provides the capability to operate under local control or unattended remote
control. Under local control, it sets the RDA mode and performs the various startup and shutdown
procedures, based on control inputs from the maintenance console. Under unattended remote
control, the same operations are performed in response to the control inputs from the MSCF. The
program generates the pedestal control commands, closes the position loops of the pedestal, and
serves as the control processor for the PSP.

2-7
2.2.8 Communications. A wideband communications link exists between the RDA and
the RPG. This link supports the synchronous, bidirectional, full duplex communications of data
between the RDA and the RPG. For additional information, see Section 2.5.
2.3 Radar Product Generator. The RPG provides real-time generation, local storage, and
distribution of WSR-88D products and base data. The RPG also includes hardware and software
required for base data ingest from the RDA, control, and status monitoring and error detection.
The RPG processes base data from the RDA and processes meteorological and
hydrometeorological products by execution of resident algorithms.
The RPG also provides the capability to control RDA operational modes and VCPs and to monitor
WSR-88D unit status. Status and error messages from the RDA and RPG are stored and any status
changes in the RDA or RPG are sent to the Associated User display system.
The major hardware components of the RPG are the computer processors, the MSCF, and
equipment to support wideband and narrowband communications.
2.3.1 Master System Control Function. The main screen of the RPG Graphical User
Interface (GUI) is the primary software application the operator uses to interact with the MSCF
terminal. While part of the function of the GUI is to provide a launch point to various system
control windows (applications), it is important to note that a great deal of significant system
information can be obtained from this screen. The vital information necessary to determine the
overall operability of the system is available at a glance. The MSCF, via the GUI, provides control
and monitoring of the RPG, the RDA, and communications.
Sets and subsets of products to be generated, distributed, and placed in mass storage can be
selected. In addition, modifications to adaptation parameters of the following functions can be
made at the MSCF:
1. Products sent in response to one-time requests from a Non-Associated User,
2. Task priority for load control at the RPG when overload conditions occur,
3. Load control across communications lines from the RPG when overload conditions
occur,
4. Thresholds for RPG usage levels that, when exceeded, result in a warning message at
the MSCF,
5. Product-alert pairing, and
6. Meteorological and hydrological algorithms.

2-8
The MSCF provides the interface with the RPG computer to perform functions such as
initialization, rebooting, startup of unit software, and various disk operations.
The different types of status information available at the RPG GUI are listed below and depicted in
Figure 2-1.
1. RDA Alarms - Whenever alarm conditions are present at the RDA, an indication
descriptive of the alarm type will appear in the upper left portion of the GUI.
2. Current Azimuth and Elevation Angle - The thick black line on the radome perimeter
shows the beginning through the current antenna azimuth for the present elevation cut.
The value in the center of the radome (along with the emanating yellow beam) shows the
antenna's current angle.
3. RDA State/Operability Status - Shows the RDA operational state and operability
status.
4. Current VCP/wx mode - Indicates which of six VCPs (11, 12, 21, 121, 31, 32) is
currently in use. The beams emanating from the radome also provides this information:
1 VCPs 11, 12 = 14 beams / elevation angles
5. Current volume start time - Indicates the date and time at which the current volume
scan began.
6. RDA Control - Indicates where control of the RDA resides (remote [i.e., RPG/MSCF]
or local [RDA/Man Machine Interface (MMI) or FAA RMS]):
1 EITHER means the RDA is in Standby and either remote or local can take
control.
2 RPG: The RDA is controlled remotely. (RPG/MSCF)
3 RDA: The RDA is controlled locally. (RDA/MMI or FAA RMS)
7. RPG State/Mode - Shows the current RPG State (Operate, Standby, and Shutdown) and
the mode in which it is operating. A state of "Transition" will be displayed during the time
period the RPG is changing from one of the three states to another. RPG mode can be
either Operate or Test.

2-9
8. RPG Alarm - Present whenever maintenance required and/or maintenance conditions
exist.
Figure 2-1
RPG Graphical User Interface
Status Indicators Within the RPG Graphical User Interface (GUI). The overlayed numbers
(in yellow) correspond to those discussed in the text.

2-10
9. Power Source - The power source is indicated by the presence of either telephone
poles/power lines (utility) or a generator shelter (indicating the generator as the power
source). If the selected power source was the generator, and utility power source
were available, a textual indication (Utility Power Available) would be displayed.
10. Wideband Status - Displays the current status of the RDA - RPG Wideband
communications link.
Also displayed is the data transmission enabled status (whether or not the data for a given
moment will be sent from the RDA to the RPG) for each of the three base moments.
1 Green means enabled.
2 Red indicates that the transmission of a particular moment or moments has

been disabled.
11. Product Distribution Comms - Indicates the current state of Product Distribution
Communications (formerly referred to as Narrowband Communications).
12. Archive II – Recorded off site.
13. Archive III – Recorded off site.
14. Precipitation Category - Shows the current Precipitation Category as either

significant, light or none.
15. Auto VAD Update - Indicates whether the Auto Velocity Azimuth Display (VAD)
Update function is turned ON or OFF.
16. Auto PRF - Indicates whether the Auto PRF function is turned ON or OFF.
17. Delta System Calibration - Displays the current Delta System Calibration value and
indicates whether or not Auto Calibration is enabled (Auto) or disabled (Manual).
18. Load Shed Status - Indicates whether or not the utilization rate of one (or more) of the
6 load shed categories has breached either the warning or alarm thresholds (alarm takes
precedence over warning).
19. Audio Alarms - Indicates whether Audio Alarms (for conditions monitored by the
RPG) are Enabled or Disabled.

2-11
20. RDA Messages - Indicates whether reception of text messages sent from the RDA is
Enabled or Disabled.
21. System Status/Alarms - The Status line displays the most recent system status change
- a corresponding entry will be submitted to, and can be displayed from, the system status
log. The Alarms line shows the most recent alarm condition that has developed within the
system. Since the occurrence of an alarm condition also represents a change in status, an
alarm condition also warrants an entry in the status line.
2.3.2 Archive II and III.
2.3.2.1 Archive II. These data, defined in Part A of this Handbook and also
known as “Level II,” are not recorded on site. These data are routed to the 4 ports on the Base Data
Distribution System (BDDS), at all DOC and DoD Contiguous United States (CONUS) sites,
where users can receive the data. The Level II data stream also contains metadata which describe
the data and the operating conditions of the radar when the data were generated. Level II data,
from all DOC and 13 DoD CONUS sites (currently), are sent electronically to the NCDC for
meeting NWS data archive requirements, National Centers for Environmental Prediction (NCEP)
for NWS real-time operational requirements, and are readily available to other users in real time.
2.3.2.2 Archive III. These products and data, defined in Part A of this Handbook
and also known as “Level III,” are no longer recorded on site, but are sent (along with specified
additional products and data) from all DOC radars to the NCDC archives via the NWS Advanced
Weather Interactive Processing System (AWIPS) and NWS Radar Product Central Collection
Dissemination Service (RPCCDS). (The list of the products collected is available at:
http://www.nws.noaa.gov/ops2/prodserv.html.) The same set of products and data are sent to the
NWS RPCCDS and the NCDC, from all DoD and DOT radars except for 4 remote overseas DoD
radars.
2.3.3 Radar Product Generator Communications Equipment. A wideband

communications link exists between the RDA and the RPG. The RPG narrowband
communications links include modem interfaces to Associated and Non-Associated Users, Other
Users, and remote or local MSCF. (For additional detail about communications, see Section 2.5.)
2.3.3.1 Wideband Communications Equipment. See Section 2.5 for details in

regard to wideband communications.
2.3.3.2 Narrowband Communications Equipment. Narrowband

communications use 2- or 4-wire, leased lines and 2-wire, switched lines. The 2- or 4-wire, leased
lines are used for communications with the MSCF, and an associated principal user display system
and between the RPG processor and a remotely located MSCF. The 2-wire, switched lines are
used for dial-up communications with Non-Associated Principal User Display Systems.

2-12
Narrowband communications equipment sends and receives selected products, alarms, and status
and command messages to and from Associated and Non-Associated Principal User Display
System and Other Users.
Narrowband communications using 4-wire leased lines are used to support Frame Relay service.
Frame Relay service is a high-performance Wide Area Network (WAN) protocol that operates at
the physical and data link layers of the Open Systems Interconnection (OSI) reference model.
Frame Relay is a packet-switched technology. Packet-switched networks enable end stations to

dynamically share the network medium and the available bandwidth. Frame Rrelay service
equipment consists of the Frame Relay Hub Router located at the NWS Weather Forecast Offices
controlling DoD and FAA radars and a Channel Service Unit/Data Service Unit (CSU/DSU)
module installed in all RPG Routers.
2.4 Associated Principal User Display Systems. The NEXRAD agencies used the PUP as the
standard WSR-88D display device when the network was first deployed. In order to meet agency-
unique WSR-88D data display and processing requirements, the NEXRAD agencies have migrated
to a new generation of display systems which are considered Associated Users, as were the PUPs.
The display capabilities and default display parameters of each system vary. Below is a summary
of the primary NEXRAD agency associated Principal User Display Systems connecting to RPGs:
2.4.1 Advanced Weather Interactive Processing System. The NWS’s AWIPS is a

distributed data processing system used at field offices, regional offices, River Forecast Centers,
national centers, and NWS Headquarters to integrate information received from all other
observational and analytical elements of the NWS Modernization. The AWIPS also provides a
nationwide communications network for distributing weather data, products, and services.
2.4.2 Weather and Radar Processor. The FAA’s Weather and Radar Processor (WARP)
is an FAA computer network that displays WSR-88D product data on air controllers’ displays in
Air Route Traffic Control Centers (ARTCCs). The WARP also collects, formats, and distributes
weather information to Center Weather Service Units at FAA ARTCCs.
2.4.3 Integrated Terminal Weather System. The FAA’s Integrated Terminal Weather
System (ITWS) provides terminal aviation system users with safety and planning products that
characterize current terminal weather situations as well as forecast about 30 minutes into the future.
The ITWS integrates data from the WSR-88D, Terminal Doppler Weather Radar (TDWR), Airport
Surveillance Radar-9 (ASR-9), Low-level Windshear Alert System (LLWAS), Automated Surface
Observing Systems (ASOS), and other NWS systems.
2.4.4 Open System Principal User Processor. The OPUP Program was started in 1996 as
a direct replacement to the PUP. The OPUP is maintained as the WSR-88D baseline display
system. The OPUP provides the capability to request products from the RPG; display, manipulate,
and locally store products; monitor the status of the RPG and RDA; and perform archival
functions. The OPUP has evolved into three separate sized OPUPs to support various missions.

2-13
The Small OPUP is a direct replacement for the legacy PUP at select Air Force and Navy
locations. It consists of a single workstation.
The Medium OPUP supports direct connections to up to seven WSR-88Ds. They are
located at centralized forecast facilities in the Pacific for the Air Force, and the East and
West coasts for the Marine Corps. The medium OPUP consists of up to three workstations
along with communications hardware and server.
The Large OPUP supports direct connections to up to 24 WSR-88Ds. They are located at
four centralized forecast facilities across the United States. The large OPUP consists of up
to 10 workstations along with communications hardware and a server.
2.5 Communications. The WSR-88D communications equipment supports wideband and

narrowband communications between the RDA and RPG and the RPG and OPUP/Associated User
display systems, respectively.
2.5.1 Wideband Communications. The wideband communications link includes Direct

Memory Access interfaces within the RDA and RPG computers, Wideband Communications
Control Modules (WCCMs), Digital Communications Units, and a communications medium.
Three types of media used for transferring wideband/base data, status commands, and controls are
listed below. Commercial T1 links, using combinations of the following media, are used at several
sites.
1. Wire - used when the RDA and the RPG are up to 122 m (400 ft) apart
2. Fiber optics – the transmitter uses either a laser or a light-emitting diode light
source, depending on the length of the link. The
WSR-88D fiber optic transceiver components and cable design can support a
fiber cable distance of up to 2 km (1.2 mi).
3. Microwave Line of Sight (MLOS) - consists of a transmitter and receiver that are
shielded against electromagnetic interference and encased in a single radio rack. A
dish antenna is used for MLOS transmission and reception. Range considerations for
this type of medium, without a repeater, are separation of the RDA and the RPG of up
to 22 nm (41 km).
Wideband communications support three types of data exchange:
1. Data transfer between processors - data are transferred between the RDA and the RPG
by direct memory access at a rate of 1.544 megabits per second over full-duplex,
synchronous lines
2. Control commands from the processors to the WCCM and link equipment - control of
the wideband communications link is executed by writing commands to the control

2-14
module. These commands make transfer of data between the computer and the link
possible
3. Status information from the communications equipment to the processors - a report is

provided when a change in status of the communications link occurs. Error detection
in the communications link is provided for by the WCCM, which resides on a circuit
board internal to the computer chassis.
2.5.2 Narrowband Communications. Narrowband communications include links to and

from the RPG and Associated Users; MSCF; and other non-Associated Users. Ports for
narrowband links are designated as dedicated, which is a full time connection; dial-up, which refers
to a dial-in, part time connection; or Frame Relay, which is a full time connection.
Frame Relay and analog service to and from the RPG are used for product dissemination,
commands, control and data request between the RPG and various associated Principal User
Display Systems. Data are transferred at various rates, depending on the location of the user
display system. The MSCF equipment is connected to the host RPG by either a X.25 dedicated
analog circuit or a Frame Relay circuit.
Most, but not all, DoD and FAA sites use Frame Relay service. Frame Relay service is not used at
all sites because it cannot be obtained at every location. Where Frame Relay service is used, the
Frame Relay Hub Router supports Frame Relay circuits that have replaced modem interfaces using
X.25 dedicated analog circuits. The Frame Relay circuits are used to send radar products to
AWIPS and to control and command DoD and FAA radars (MSCF data). Frame Relay service
also supports One Time Requests (OTR) from DoD and FAA radars via the AWIPS WAN.

2-15
2-16
CHAPTER 3
UNIT FUNCTIONAL FLOW
3.1 Introduction. The purpose of this chapter is to describe the flow of data through the
application software of the major functional areas of the WSR-88D unit to enable the user to better
understand and manage the available data resources. Emphasis is placed on the flow in the RPG.
3.2 Radar Data Acquisition. The RDASC program is used to control the real-time operation of
components of the RDA. The RDASC program monitors and assesses the performance of the
RDA, initiates automatic calibration, performs calibration calculations, and reports RDA status to
the RPG functional area. It formats reflectivity, mean radial velocity, and (velocity) spectrum
width data input from the signal processing subsystem, attaches header data, and initiates data
transfer to the RPG. The RDASC program is partitioned into the following functional capabilities
(Figure 3-1):
• Handle Maintenance Console,

• Monitor and Calibrate RDA,
• Form RDA State,
• Handle Wideband,
• Control RDA,
• Control Signal Processor, and
• Control Pedestal.
3.2.1 Handle Maintenance Console. This function provides the RDASC Program
interface with the Maintenance Console Keyboard and the Data Acquisition Unit (DAU). Inputs to
the Handle Maintenance Console function are received from the Keyboard, Monitor, Calibrate
RDA function, the DAU, and the Control RDA function.
Command entries from the keyboard are validated to ensure that they conform to the allowable set
in format and type. Associated parameter entries are validated to determine if they are within the
allowable range. When the RDA is in the remote control state, certain commands can only be
issued from the RPG. Invalid entries result in error messages being displayed; valid entries are
acknowledged on the display. Command and parameter entries are made through the use of menus
that prompt the user for inputs.

3-1
Handle Handle Control Monitor & Control
Maint Wideband Pedestal Calibrate Signal
Console RDA Processor
-Console -RPG -Angle -Status -Data

Commands Commands State Collec-
-Control -Cut tion
Data Complete Complete
-Data
Available
Control Data Stores

Pedestal
-Initialize
-Control Data
State -Status
-Volume -State Data
Coverage -Clutter Map
Pattern -Scan Profile
–State Data
Form RDA Control Control Monitor & Handle

State Signal Pedestal Calibrate Maint
Processor RDA Console
-Specify -PSP Mode -Profile -Status -Commands

data to Param Instruc- Messages to DAU
be sent & -Profile tions & Alarms
archived Instruct
at Level -SPS Data
III Pointers
Handle
Wideband
-Control RPG Other

Messages -Data
Figure 3-1
Radar Data Acquisition Status and Control Program Data Flow

3-2
An Enter Log Data command provides for the entry of data into a maintenance log file. The entry
procedure includes prompts that assure that complete information is input for each event to be
logged. A Display Log Data command provides for the display of previously entered data. The
capacity of the log is 100 records with an overload warning issued when the log contains 90 or
more.
An Inspect/Change RDA Adaptation Data command allows the user to inspect or change
adaptation data. Changes to certain data are only allowed under password protection with two
levels of protection provided. (A listing and description of site-adaptable parameters applicable to
the RDASC are contained in and specified by WSR-88D Operator Handbook, Guidance on
Adaptable Parameters, Volume 3, RDA. The handbook can be obtained at:
http://www.roc.noaa.gov/ssb/sysdoc/Operations.asp. The authority to change these parameters is
addressed in Part A of this Handbook.) When adaptation data are changed, the original version of
the data is retained with changes used to generate a current version. Both versions are maintained
on disk and the capability to edit either version is provided.
Summary RDA status and alarm messages, as designated by the Monitor and Calibrate RDA
function, are displayed in dedicated and protected areas of the alphanumeric screen. These
displays are active at the same time that menus and other data are displayed. The status display is
updated automatically when changes in status occur. Performance data can be displayed upon
command with separate menus provided for the various categories of data.
When a Cold Startup or Restart command is issued from the Control RDA function, actions are
taken automatically to initialize the DAU and Maintenance Console function. When the DAU and
Maintenance Console interfaces are initialized, the DAU initialization status and the Maintenance
Console initialization status contained in the RDA performance data are set to OK or Fail,
depending on whether or not the initialization was successful. If an error occurs during the
initialization process, appropriate alarms are set.
Outputs from this function are directed to the Maintenance Console and its alphanumeric display,
to the RDA Data function, to the Data Store, and to the Control RDA function.
3.2.2 Monitor and Calibrate Radar Data Acquisition. This function generates RDA
performance data based on Built-In Test (BIT) data and other inputs from the RDA equipment,
other program functions, and the Operating System. Inputs consist of pedestal angle state
information from Control Pedestal function and signal processor, transmitter, receiver, rf generator
BIT data, and tower and utilities data from the Data Store. Data Store also provides scan
information for scheduling tests and calibration data from the signal processor.
The function maintains the RDA summary status at the Maintenance Console for transmission to
the RPG through the Form RDA Data function, whenever changes in the data occur and when
RDA status is requested by the RPG. The RDA performance data are also formed at the

3-3
Maintenance Console for transmission to the RPG as they are generated by this and other functions
of the RDASC program.
3.2.2.1 Evaluate Hardware Status. The RDA performance data are generated for
each hardware configuration item based on inputs from the DAU, pedestal, and receiver and signal
processor. The DAU BIT inputs are available from the Handle Maintenance Console function.
Pedestal and receiver and signal processor data are received from the Control Pedestal and Control
Signal Processor functions. Calibration data are received from the Perform RDA Calibration
subfunction. RDA alarms are set when discrete BIT data indicate a fault or when analog BIT data
or calibration data are out of tolerance.
3.2.2.2 File Maintenance. Whenever data are read from a disk file or written to a
disk file, the disk file read/write status is updated based on the status reported by the Operating
System; it is then included in the RDA performance data. Disk files are maintained for the RDA
state file, bypass map, calibration data, adaptation data, censor zone, and remote VCP.
3.2.2.3 Perform Radar Data Acquisition Calibration. Calculations of

calibration parameters are performed when RDA calibration processing is scheduled by the Control
RDA function. This activity generates applicable RDA performance data, performs calibration
checks, and generates alarms, when appropriate. It is performed during the period of elevation
transition between volume scans or elevation cuts and at startup.
Outputs from the Monitor and Calibrate RDA function (summary status, performance data, alarms,
and calibration parameters) are sent to the Handle Maintenance Console and Control Signal
Processor functions.
3.2.3 Form Radar Data Acquisition Data. This function assembles data to be output on
the wideband communications links, inserts required header data, generates the associated data
parameter blocks, and passes the data parameter block to the Control Wideband function for data
transmission. Data are input from the Control Signal Processor, Monitor and Calibrate RDA,
Control RDA, and Control Pedestal functions.
When notified of data to be output on the wideband data links, the Form RDA Data function
generates a header for the data, which includes message size, message type, an identification
sequence number, a date and time tag, the number of separate message segments that make up the
data message, and the individual segment number. Other header information for base data includes
pedestal angle positions and a calibration parameter that is derived from automatic calibration
routines or RPG input. Outputs include a parameter block for each data set to be transmitted,
header data segments to be incorporated into the data sets, and notification to the Handle Wideband
function that data are available.
3.2.4 Handle Wideband Data. This function supports and controls the wideband
communications link to the RPG. Inputs include RPG to RDA data received over the wideband

3-4
link, header data, and an RDA data pointer list from the Form RDA Data function, message data
from data stores, wideband state from the wideband communications handler, and wideband
control from the Control RDA function. This function is divided into three subfunctions: Handle
Wideband Link, Transmit Wideband Data, and Receive Wideband Data.
3.2.4.1 Handle Wideband Link. This subfunction provides an interface with the
channel terminal manager, which is part of the Operating System. It provides the capabilities to
establish and terminate communication links and to control transfer of data between the link
interfaces and other functions of the RDASC program.
3.2.4.2 Transmit Wideband Data. When notified by the Form RDA Data
function that radials of data are ready to be transmitted, this subfunction determines whether
sufficient buffers are available for the Control Signal Processor Subsystem (SPS) function to
continue reading data out of the PSP. If sufficient buffers are not available, it determines which
interface (RPG, wideband user, or Archive II) is responsible for the unavailability of sufficient
buffers and deactivates it to prevent it from interfering with data transmission over the other
interfaces.
3.2.4.3 Receive Wideband Data. When notified by the Handle Wideband Link
subfunction that a message has been received, this subfunction notifies the Control RDA function
that data are available.
The Handle Wideband function outputs include RPG to RDA data received over the wideband link,
RDA to RPG data and RDA to user data to be transmitted over the wideband link, wideband status
data from the wideband communications handler, and wideband control from the Control RDA
function.
3.2.5 Control Radar Data Acquisition. The Control RDA function performs processing
in response to control commands entered at the RDA Maintenance Console or received from the
RPG through the Handle Wideband function. It also provides overall sequence control of the
volume coverage by initiation of processing within the Control Signal Processor and Control
Pedestal functions based on the selected VCP. Inputs include entries from the RDA Maintenance
Console and the MSCF, hardware status information from the hardware interfaces or other RDASC
functions, and control data from the RPG. This function is further divided into two subfunctions:
3.2.5.1 Control Radar Data Acquisition State. The RDA states are
combinations of control states and operating states. Control states include local (RDA), remote
(RPG), and either (RDA or RPG). Operating states include standby, restart, startup, playback,
offline operate, and operate. Control states define the commands that can be accepted from the
RDA Maintenance Console or RPG while operating states define the processing that is performed
by the RDA in response to commands. Transitions between the various control states and
operating states can take place due to commands from the RDA or RPG, or automatically when
specified conditions occur.

3-5
3.2.5.2 Volume Coverage Control. The Volume Coverage Control subfunction initiates
and controls the execution of the selected VCP as well as the RDA calibration functions. The
selected VCP is determined by the command issued at the RDA Maintenance Console or the RPG.
If no command was issued, the default VCP identified in the RDA state file is selected. If a new
VCP is selected while a volume scan sequence is in progress, the new VCP takes effect when the
next volume scan starts.
Outputs of the Control RDA State include status messages and alarms to the Handle Maintenance
function; status data, commands, and parameters to other RDASC program functions; and
transmitter control commands.
3.2.6 Control Signal Processor. This function loads the SPS controls into the SPS and
receives and unpacks the SPS outputs. The SPS input data consist of base data and surveillance
power data, calibration data, on-line performance monitoring data, signal processor loads data, and
control RDA function inputs.
When power is first applied to the SPS at startup or restart following a power loss, microcode
required by the PSP is automatically downloaded to the PSP by the SPS. This download may also
be commanded from the RDASC processor.
3.2.6.1 Clutter Filter Data Transfer. During normal radial processing, the proper
radial of the Clutter Suppression Map is downloaded to the SPS. The elevation dimension of the
map data is taken from the elevation angle for the cut in the VCP. The azimuth dimension of the
map data is the pedestal angle projected forward to account for pipeline delays. The Clutter
Suppression Map is derived at startup and whenever a new set of Clutter Suppression Regions are
received by the RDA (downloaded by the MSCF). This map is derived by merging the default
suppression levels defined by the Default Notch Width Map and clutter locations contained in the
Bypass Map definition with the operator-defined Clutter Suppression Regions. When compiling
the Clutter Suppression Map, operator-defined Clutter Suppression Regions take precedence for the
application of clutter suppression; however, the Bypass Map and Default Notch Width Map control
the application of suppression for every range bin NOT contained within an operator-defined
region. The actual notch width values assigned to each radial of the Clutter Suppression Map are
determined by the notch width verses scan rate and suppression level table maintained in RDA
adaptation data, where the scan rate is determined by the VCP and the suppression level is
determined from the level defined in the Clutter Suppression Map.
3.2.6.2 Compute Interference Detection Rate. When the Interference Detection

Unit is configured, the function computes the interference detection rate contained in the Summary
RDA Status.
Outputs to the RDASC resident portion of the Signal Processing Program consist of macrocode
setup parameters. These parameters are transferred either by shared memory or in the form of

3-6
subroutine parameters. Receiver and signal processor control and supplemental data are
downloaded to the SPS through Direct Memory Access transfer.
Output from Data Stores includes on-line performance monitoring data from the SPS, receiver, and
RF generator, calibration data used by the Monitor and Calibrate RDA function, and digital radar
data that are made available to the Handle Wideband function.
Other outputs are digital radar data buffer pointer consisting of address information needed to
access digital radar data that are sent to the Form RDA Data function; a calibration parameter used
by the PSP to calculate reflectivity from signal power, which goes to the Form RDA Data function;
notification to the Control RDA function indicating that the SPS data report has been received and
the status of the SPS; and a flag used to initiate transition to the next elevation angle scan, which is
sent to the Volume Coverage Control subfunction.
3.2.7 Control Pedestal. The Control Pedestal function of the RDASC program provides
three basic and essentially separate subfunctions. The primary function is to perform closed loop
servo control of the pedestal azimuth and elevation positions. This control may be by either rate or
position. Scan Profile Data input to the Control Pedestal function provides the basic information
describing the desired volume scan scenario. The Control RDA function instructs the Control
Pedestal function as to which set of data is to be used for a particular volume scan. It also provides
real-time commands to commence or terminate a volume scan or, when appropriate, to restart the
scan at any particular point after a short-term power outage. In return, Control Pedestal informs
Control RDA when particular angle or angle rate transition sequences are stabilized so that radar
data processing can be initiated.
A secondary function of Control Pedestal is to provide azimuth and elevation angular positions to
other functions. These angles are appended to Digital Radar Data by the Form RDA Data function.
The remaining function of Control Pedestal is to receive pedestal built in test equipment (BITE)
data from the pedestal and forward these data to the Monitor and Calibrate RDA function.
Inputs include scan profile data that define the information needed to perform various volume scan
scenarios; pedestal data, which include pedestal azimuth and elevation angular positions; and
angular rates as well as pedestal BITE data. Profile instruction input is provided from the Control
RDA function as well as pedestal position correction parameters.
When a startup command is issued by the Control RDA function, pedestal self tests are conducted
by issuing appropriate commands to the pedestal and checking the test results.
3.2.7.1 Generate Angular Position. This subfunction generates the azimuth and
elevation position commands for the pedestal as a function of time. The scan patterns are specified
in terms of constant rate or position segments, or both, separated by transition events. Operational
scan strategies specified to date require only constant elevation continuous azimuth rotation

3-7
patterns. For this type of pattern, the segments are constant elevation positions with a constant rate
azimuth rotation and the transition event is the start of a new cut.
3.2.7.2 Control Pedestal Servo. The position of the pedestal in azimuth and
elevation is controlled by this subfunction. It compares present pedestal position to commanded
pedestal position and generates rate commands to the pedestal based upon position error and
desired rate.
3.2.7.3 Handle Pedestal Input and Output. Communications with the pedestal
in support of all pedestal control functions is provided by the Handle Pedestal Input and Output.
These communications are implemented over a serial data link through a controller tied to the
RDASC computer.
3.2.7.4 Format Pedestal Built-In Test Equipment. The Format Pedestal BITE
continually updates pedestal state data, which can be accessed by the Monitor and Calibrate RDA
function for purposes of ascertaining the operability state of the pedestal.
Outputs of the Control Pedestal function include pedestal control commands, consisting of
commands regarding the azimuth and elevation drive rates and pedestal data requests sent to the
pedestal through the Handle Pedestal Input/Output (I/O); angle state, consisting of data and flags
that are sent to the Control RDA function; and pedestal state, consisting of pedestal BITE data and
a fault flag that are accessed by the Monitor and Calibrate RDA function.
3.3 Radar Product Generation. The RPG functional area controls the operation of the entire
WSR-88D unit. It is organized into modules that, taken together, accomplish the following
functions:
• System Status and Control,

• Base Data Acquisition,
• Preprocess Base Data,
• Product Generation, and
• Product Distribution.
Figure 3-2 contains a summary of the flow and functionality of each of these modules. The
purpose of these modules is to transform digitized weather data received from the RDA,
supplemental external data, and user input data into base products, derived products, alphanumeric
products, and data arrays in forms suitable for archive and distribute to WSR-88D users.

3-8
RDA RPG EXTERNAL
USERS
-Status -HCI
-Performance Commands -Non-
Data -Adaptation Associated
-Radial Data Data -Associated
System Control
-RDA Control
-Product Generation Control
-Product Distribution Control
Product Product Control Base Data

Distribution Generation Scheduling Acquisition
-Routinely -Directly -Routinely -Routinely

Scheduled from Base Scheduled Scheduled
-One-Time Data -One-Time -One-Time
-Alert-Paired -Through -Alert-Paired -Alert-Paired
-Status Meteorolog- -Status -Status
-Alerts ical -Alerts -Alerts
Algorithms
Preprocess Base Data
-Data Checks
-Velocity Dealiasing
-Combine Reflectivity &
Velocity-Low Elevations
EXTERNAL Temporary
USERS Storage
(Disk)
Figure 3-2
Radar Product Generation Data Flow

3-9
3.3.1 System Control. Within the RPG, the System Status and Control function accepts
input data from the RDA, the RPG itself, and display systems. Site-adaptable parameters resident
within the RPG are listed in the WSR-88D Operator Handbook, Guidance on Adaptable
Parameters. As mentioned earlier, authority for change of these parameters is addressed in Part A
of this Handbook. These data are used for both RDA control, RPG control, and product generation
and distribution control.
The System Status and Control function also presents to the system user information concerning
overall system health through a combination of RDA and RPG status displays and RDA
Performance Data.
3.3.1.1 System Status and RDA Control. The RPG user controls the RDA
through the use of RDA Control Commands available at the RPG Human Computer Interface
(HCI). Operationally, the most frequently used commands are those specifying VCP, PRFs used
within the VCP, and Clutter Suppression Region definitions.
Other commands available allow the system user to restart the VCP, change between utility and
backup power source, place the RDA in Standby or Operate State, restart RDA application
software, and change the RDA control state between Local RDA Control and Remote RDA
Control. The RPG HCI user can only control the RDA when the RDA is in Remote RDA control.
The RPG HCI displays RDA status information to enable the HCI user to monitor the RDA for
alarm messages and performance quality. Data available to the user includes transmitter,
receiver/signal processor, antenna/pedestal, and communications performance.
3.3.1.2 Product Generation. At the beginning of each volume scan, the RPG
creates a table of products that will be produced for that volume scan. This table is made up of
products specified at the MSCF on the Generation and Distribution Control Menu, products in
response to alert thresholds being exceeded, Routine Product Set (RPS) lists input from Associated
User display systems, and one-time requests from Associated User display systems.
3.3.1.3 Product Distribution. Distribution of products is controlled according to

inputs from User display systems and the MSCF as described in Section 3.3.5.
3.3.1.4 Product Generation Control. Generation of products is controlled

according to inputs from External User Systems and the RPG HCI as described in Section 3.3.4.
3.3.1.5 Product Distribution Control. Distribution of products is controlled

according to inputs from External User Systems and the RPG HCI as described in Section 3.3.5.
3.3.2 Base Data Acquisition. The Base Data Acquisition function is responsible for
managing the wideband communications link to the RDA and processing incoming and outgoing
RDA/RPG interface control messages. Messages passed from the RDA to RPG and processed by

3-10
the Base Data Acquisition function include Digital Radar Data, RDA Status Data,
Performance/Maintenance Data, Console Message, Clutter Filter Bypass Map, and Clutter Filter
Notchwidth Map. Messages passed from the RPG to the RDA and processed by the Base Data
Acquisition function include RDA Control Commands, Volume Coverage Pattern, Clutter Censor
Zone Data, and Console Message.
3.3.2.1 RDA and RPG Radial Loadshedding. In the RPG, there are two queues
where radial data is stored and monitored. In one of the queues, radial data in RDA/RPG ICD
format (i.e., Archive II/Level II format) is buffered. The other queue contains the radial data after
conversion from interface control format to internal RPG format (this latter format is for RPG
algorithm processing convenience). Thus, the two queues represent different locations within the
radial processing stream of the RPG. Figure 3-3 contains a summary of RPG loadshedding.
The current utilization values presented for these categories are the ratio of number of unread (i.e.,
unprocessed) radials in each queue to the maximum number of radials each queue can hold. Unlike
the other loadshed categories, if the Alarm Levels are exceeded, no actual loadshedding mechanism
is invoked unless queue overflow occurs. A loadshed Alarm merely indicates the processing of
radial messages is severely backlogged and product generation based on radial data input may fail.
Product generation failures will occur if either of the queues overflows because newly arriving
radial messages will replace unread radial messages in the queues.
3.3.3 Preprocess Base Data. The RPG performs certain preprocessing functions on the
base data before making these data available for meteorological algorithm processing and product
preparation. This preprocessing includes: performing angle calculations, ensuring angle changes
are within tolerance, velocity dealiasing, and other continuity considerations. The processed
radials are then released to downstream consumers. In addition, these processed radials are saved
in a database so that one-time requests for products not routinely produced can be handled within a
reasonable time period.
3.3.3.1 Velocity Dealiasing. Velocity dealiasing is performed using two methods,

the first uses an algorithm (Part C of this Handbook) in the RPG in order to take advantage of the
processing power there and facilitate the implementation of improved dealiasing algorithms in the
future. The MPDA is also executed at this time in the RPG. We will begin with the original
method. The following description is intended only to present the main functional elements of the
complex algorithm.

3-11
RDA RPG EXTERNAL
USERS
-Status -HCI
-Performance Commands -Non-
Data -Adaptation Associated
-Radial Data Data -Associated
System Control
-RDA Control
-Product Generation Control
-Product Distribution Control
Narrowband RPG Radial

Load Shedding /RDA Radial
When products If Alarm

reside in Threshold is
narrowband exceeded. Base
queue longer Data arrives
than the Load faster than it
Shed Alarm can be processed
Threshold. by RPG.
Preprocess Base Data
-Data Checks
-Velocity Dealiasing
-Combine Reflectivity &
Velocity-Low Elevations
EXTERNAL Temporary
USERS Storage
(Disk)
Figure 3-3
Types of Load Shedding at the Radar Product Generator

3-12
The algorithm dealiases one radial of velocity data at a time, operating from the first range bin out
to 124 nm. Each velocity value is checked for spatial continuity both with previous bins, which
have been dealiased on the same radial, and with already dealiased bins near the same range on the
previous radial. The algorithm allows for large changes in velocity values if the changes occur
over limited distances. This provides the ability to avoid falsely dealiasing the velocity values
associated with meteorological features such as Mesocyclones and Tornado Vortex Signatures. If
spatial continuity checks indicate an unacceptable velocity value and the value is spatially isolated,
the algorithm uses environmental wind information (entered at the HCI and updated by the
Velocity Azimuth Display Algorithm) to attempt to dealias the suspect value.
If neither spatial continuity nor environmental wind checks enable the dealiasing of the suspect
value, the value is rejected. Rejected velocity values are temporarily stored for possible reinsertion
into the radial. If five velocity values in a row are rejected, they will be reinserted into the radial.
If a radial cannot be dealiased completely, it is not used for azimuthal continuity checks on the next
radial's velocity values. In this case, the previous completely dealiased radial continues to be used
for the azimuthal continuity checks. If five successive radials cannot be dealiased completely, no
further azimuthal continuity checks are made during the rest of that scan's dealiasing process. This
feature prevents unsuccessfully dealiased areas from propagating azimuthally around the scan.
The second method uses the MPDA approach. The MPDA dealiasing scheme is a multi-step
process that arrives at a final dealiased velocity solution at each radar gate. Throughout the
processing steps described below, “seed” velocities are used to check the gate-by-gate MPDA
results for consistency. These seeds can be single previously dealiased velocity solutions along the
same radial, averages of previously dealiased gates, or estimates from the Environmental Wind
Table (EWT).
The main dealiasing steps are:
a) Solutions from Velocity Triplets (tight constraint)
This first step considers only gates at which three velocity estimates are present. These
estimates for the same gate are obtained by sampling with three different PRFs. The three
estimates must be dealiased within a small velocity difference of each other and within a
threshold velocity difference of a seed velocity as previously defined for a final solution to
be accepted. In general, this processing accounts for about 57% of the final dealiased field.
b) Solutions from Velocity Triplets (relaxed constraint)
The second step considers only gates at which three velocity estimates are present. The
three estimates and a seed velocity must be dealiased within a larger velocity difference
than step a) of each other.
c) Solutions from Pairs within Triplets

3-13
The algorithm next attempts to dealias velocity pairs that did not pass the dealiasing tests in
steps a and b. The use of velocity pairs may be due to there being only two estimates at a
particular location in space or due to the failure to find a solution using triplets. In the case
where triplets are present three solutions are possible. However, the first pair that provides
an acceptable solution is retained. Note that at 4.3° elevation, where only two velocity cuts
are obtained, MPDA processing begins with this step.
d) Solutions from Single Estimates
At this point the remaining solutions are derived from the single velocity estimates that
exist within the unsolved triplets and pairs, and those locations in space that only contained
one estimate. The single estimates are dealiased using seed values from the previous steps
and increasingly relaxed thresholds. Once this processing is complete, more than 99.99%
of the gates contain a final dealiased velocity value.
e) Use original velocity estimates
The remaining gates are assigned one of the three original velocity estimates that are
closest to an average of the surrounding dealiased gates.
Error Mitigation Schemes
After steps a through d described above are complete, error mitigation is applied to check
for outliers and for azimuthal and radial inconsistencies in the dealiased field.
a) Despeckling
A despeckling function is applied after each processing step. This routine checks for single
velocity gates whose solutions differ significantly from surrounding gates. Several
averages of surrounding gates are checked against the gate in question. If the gate can be
dealiased within a strict threshold of one of the averages, it is assigned a value. Otherwise,
it is set to missing and waits further processing.
b) Azimuthal Error Correction
This routine searches for runs of gates along radials that differ significantly when
compared to adjacent azimuthal values. Adjacent azimuths on both sides of the azimuth in
question are considered in the checking. If the azimuth in question can be dealiased within
a strict threshold of its adjacent azimuths, values are assigned to it. Otherwise, the gates on
the radial are set to missing and processed further.
c) Radial Error Correction
This routine searches for large gate-to-gate jumps along radials. If a jump is encountered,
an attempt is made to dealias it into the correct Nyquist interval based on averages of other
radially adjacent gates within a strict threshold. If the gates in question can be dealiased,
new values are assigned, otherwise the gates are set to missing and wait further processing.

3-14
3.3.3.2 Combining Reflectivity and Doppler Data. To help achieve optimal
clutter filter performance, the lowest elevation angles of a VCP (usually 0.5° and 1.5°) are
designated split cuts. That is, the RDA collects reflectivity (also called surveillance) data and
Doppler data on separate scans. The RPG combines these surveillance and Doppler data into a
format consistent with that for higher elevation scans. For a given Doppler data radial, the
combining process involves selecting data from the surveillance radial closest in azimuth position
to the azimuth position of the Doppler radial.
3.3.4 Product Generation. The RPG Product Generation function coordinates the
generation of routine and one-time requested products. Routine requests include products to be
generated by default as well as those products requested on a routine basis from External Users via
RPS lists. One-time requests (OTRs) include products requested on a one-time basis via OTR
messages from External Users, as well as those products generated based on an Alerting criteria
being met (Section 3.3.4.2.1). The RPG user can control which products are generated by default
by editing the RPG Product Generation Table at the RPG HCI.
3.3.4.1 Routine Product Scheduling. The Routine Product Scheduler is aware of

algorithm and product dependencies based on information defined in two configuration tables: the
Product Attributes Table (PAT) and the Task Attributes Table (TAT). The PAT lists the types and
attributes of data or products the RPG can produce. Attributes include dependent data type(s), the
data producer (algorithm), and data-dependent parameters. The TAT lists for each RPG algorithm
the data type(s) input to the algorithm and the data type(s) produced by the algorithm.
For each product request, the Routine Product Scheduler uses information contained in the PAT
and TAT to determine what data types must be produced in order to generate the requested product.
As an example, if Algorithm 1 ingests data type A and produces data type B, and Algorithm 2
ingests data type B and produces product C, then a request for product C requires data types A and
B be produced. Hence, a request for product C results in additional requests for data types A and
B.
At the beginning of each volume scan, the Routine Product Scheduler builds a Master Request List
composed of all algorithm outputs (data types or products) which must be generated in order to
satisfy default product requests, External User product requests, unsatisfied one-time product
requests, and the Alerting function. After the Master Request List is built, the list is published by
the Routine Product Scheduler for all algorithms to read.
The RPG uses a data-driven model for algorithm processing. During the course of a volume scan,
every algorithm in the RPG periodically polls whether its input data has been generated for that
volume scan. If the input data has been produced, the algorithm then checks the Master Request
List for requests for one or more of the algorithm’s outputs. If input is available and output is
requested, the algorithm processes the input data to produce the output data. Once produced, the
output data is written to either an intermediate data store (to be read by down-stream consumers),

3-15
the product database if the output is a product or the data is to be used to support one-time requests,
or in the case of radial data, stored in the Replay Database.
3.3.4.2 One-Time Product Scheduling. When the date and time field is left blank
in a one-time request, the Product Generation Function queries the product database for the product
in the current volume scan. If a match is found, information about the product is sent to the
Product Distribution function (Section 3.3.5). If no match is found, the Product Generation
Function queries the product database for the product in the previous volume scan. If a match is
found, information about the product is sent to the Product Distribution function. Otherwise, the
Replay List is checked to see if the product can be generated on a one-time basis. For those one-
time requests that can be generated, the RPG checks whether the product is to be generated from
the Replay Database or from warehoused intermediate data. The product is then scheduled for
generation through a rerun of base data or warehoused intermediate data. If the one-time requested
product can not be generated, the product request is scheduled for generation during the next
volume scan as a routine product request.
3.3.4.2.1 Alert Processing. Whenever an Associated External User establishes

connection to the RPG, the external system automatically uploads to the RPG the Alert Request
message. This message defines the locations and categories for which the External User wishes to
be alerted. The External User can modify and resend the Alert Request Message at any time during
connection to the RPG.
During the course of a volume scan, the Alert Processing function ingests data generated by RPG
algorithms. These data contain the necessary information for alert checking of all alert categories
from all Associated External Systems. A check is made for alert conditions for each requested
category corresponding to the presently available data. If an alert condition is encountered, an
Alert Message is sent to the appropriate external system. If the External User requested that the
alert-paired product associated with the alert be sent upon detection of the alert condition, the
Alerting Processing function issues a one-time request on behalf of the External User for the alert-
paired product. By requirement, only products that are or can be generated using data from the
volume scan in which the alert occurred can be paired to an alert.

3-16
3.3.4.2.2 Replay Data Products. Product requests (one-time and alert-paired) that
are not already generated, and are too late to be processed on the current stream of base (radial)
data, may be generated by a replay of radial data or warehoused intermediate data. Due to the
enormous amount of processing that would be required, not all volume-based products are included
on the replay list. The following product-generating tasks make up the current Replay List (refer to
Figure 3-4 (a through g) for task-product relationships):
Mean Radial Velocity 8-bit Velocity

Combined Shear Radar Coded Message
Reflectivity 8-bit Reflectivity
Composite Reflectivity Grid Severe Weather Analysis
Spectrum Width User Selectable Layer Reflectivity
Cross Section Storm Relative Mean Radial Velocity
Velocity Azimuth Display
If a one-time requested product is not available and the product is on the Replay List, the product
will be produced using replay data. Depending on whether the generating task ingests radial data
or intermediate product data determines which database is accessed for the information. Product
generating tasks requiring radial data, access the radial replay database. Products requiring
intermediate data access the database warehousing intermediate data.
3.3.4.3 Data Flow in RPG. Figure 3-4 (a through g) depicts the flow of data for
the generation of products, other than the User Alert Message, the Combined Attributes Table, and
the Radar Coded Message, which are composed of the outputs of other products, as indicated in
Section 3.3.4.3.3. For a description of the individual outputs, refer to Part C of this Handbook.
3.3.4.3.1 Products Produced Directly from Base Data. A list of products

produced directly from preprocessed radial data along with a summary of the data processing
follows. These products do not require processing by meteorological algorithms prior to
formatting for distribution. One exception is the Severe Weather Analysis product, which receives
input from the Shear algorithm to construct the map for radial shear data.
Reflectivity. This task obtains buffers for the six Reflectivity product types
(depending on the types requiring generation). Radial data that are input to the task are run-length
encoded for the 1 km (0.54 nm) x 1° version. For 2 km (1.1 nm) x 1° versions, the radial is
compacted by saving the largest value of two consecutive bins, after which the radial is run-length
encoded. For 4 km (2.2 nm) x 1° versions, the radial is compacted twice prior to run-length
encoding. The task processes the 1 km (0.54 nm) products first, the 2 km (1.1 nm) next, and the 4
km (2.2 nm) last. These products can be produced for any or all elevation cuts within a VCP.

3-17
Base Data
Velocity Tornado Tornado

Azimuth Detection Detection
Display Algorithm Algorithm
(VAD) (TDA 1D) (Elevation-
Based)
VAD Wind
Profile Storm Cell
(VWP) Tornado Tracking &
VAD Tornado Forecast
Detection
Vortex
Algorithm
Signature
(TDA 2D&3D)
Rapid Update
(TRU) Mesocyclone
Tracking/
Trending
Tornado Vortex
Signature
(TVS)
Figure 3-4a
Flow of Data for Generation of Products

3-18
Base Data
Mesocyclone 1D/2D Mesocyclone

Detection 1D
Mesocyclone 3D Mesocyclone
Detection 2D
Storm Cell
Mesocyclone Mesocyclone
Tracking &
Product(M) Detection 3D
Forecast
Mesocyclone
Tracking/
Mesocyclone Trending
Rapid Update
Product
(MRU)
Figure 3-4b
Flow of Data for Generation of Products (Continued)

3-19
Base Data
Storm Cell Combined Shear

Segments Product(CS)
16-Level
Storm Cell
Components/
Centroids
Storm Tracking
Hail Index Information
Product Product (STI)
(HI)
Storm Relative
Storm Structure
Velocity
Product (SS)
Storm Relative
Mean Radial
Velocity
(Region)
Product
(SRR)
16-Level
0.27 n mi
Storm Relative
Mean Radial
Velocity (Map)
Product (SRM)
16-Level
0.54 n mi
Figure 3-4c

3-20
Base Data
Super Ob User Selectable

Layer Composite
Product (SO) Layer Composite
Reflectivity
Reflectivity
(Maximum)
(Maximum) Product
Product
(ULR)
(LRM)
Layer 1
Layer 2
Layer 3
Figure 3-4d

3-21
Base Data
Base Base Mean Radial Base Spectrum Digital Base

Reflectivity Reflectivity Velocity Velocity Width Velocity
Product (R) Data Array Product (V) Data Array Product (SW) Product
Product (DR) Product (DV) (DBV)
8-level 8-level 8-level
0.54 n mi 256-level 0.13 n mi 256-level 0.13 n mi 256-level
0.54 n mi 0.13 n mi 0.54 n mi
1.1 n mi 0.27 n mi 0.27 n mi

2.2 n mi 0.54 n mi 0.54 n mi
16-level 16-level
0.54 n mi 0.13 n mi
16-level 16-level
1.1 n mi 0.27 n mi
16-level 16-level
2.2 n mi 0.54 n mi
Figure 3-4e

3-22
Figure 3-4f

3-23
Figure 3-4g
Flow of Data for Generation of Products (Concluded)

3-24
Mean Radial Velocity and Spectrum Width. These tasks obtain buffers for the
required number of products to be generated (up to six each). The task processes only the number
of bins required for the range (0.25 km (0.13 nm) x 1° resolution needs ~ 59 km (32 nm) of bins).
For the 0.25 km (0.13 nm) x 1° versions, every bin of data is included in the product. For the 0.5
km (0.27) nm x 1° and the 1 km (0.54 nm) x 1° versions, only every second and fourth bin,
respectively, are used for product generation. The only other data processing is run length
encoding the bins of data. These products can be produced for any or all elevation cuts within a
VCP.
Composite Reflectivity. The Composite Reflectivity polar grid task determines the
polar grid coordinates for each radial and for each bin within the radial. The grid value is replaced
if a higher value is determined at a higher elevation scan corresponding to the grid value in
question. After the grid is completed, the Composite Reflectivity data are mapped from polar grid
coordinates to Cartesian grid coordinates. When this process is completed, the Composite
Reflectivity Cartesian grid task determines which holes obtained during grid conversion need to be
filled. For each hole, a search of neighboring bins is used. If at least two adjacent bins contain
data values above threshold, the hole is filled with the average of these values. The final process is
run-length encoding of the data for distribution.
Severe Weather Analysis. Radial base data and radial shear data (obtained from
the High Resolution Shear Algorithm Output) are input to the task. The only processing is run-
length encoding of the data within the defined window. These products can be produced for any or
all elevation cuts within a VCP.
Vertical Cross Section. This task builds a Cartesian map of radial data bins
required for the points selected. At the end of the volume scan, holes that have data above and
below, and are not range folded or below threshold, are filled by interpolating downward. The data
are then run-length encoded for distribution.
8-bit Reflectivity. This task builds a reflectivity product formatted as a data array
with the same spatial resolution and same data quantization as the input base data. This product
can be produced for any or all elevation cuts within a VCP.
8-bit Velocity. This task builds a Mean Radial Velocity product formatted as a
data array with the same spatial resolution and same data quantization as the input base data. This
product can be produced for any or all elevation cut within a VCP.
3.3.4.3.2 Products Produced from Hydrometeorological Algorithms. The

remainder of the products are generated from either radial data and algorithm outputs or from the
Output of other algorithms. Details of each algorithm's processing are contained in Part C of this
Handbook. All algorithms and tasks implemented in the RPG have an Output. In some cases,
these outputs may be in the form of a product. In other cases, the Output is used as an input to
another algorithm or task, such as the storm segments output, which is the sole input to the Storm

3-25
Components Algorithm. Other algorithms and tasks, such as hail and storm structure, produce
outputs for products along with storm attribute information for use by the Mesocyclone Detection
Algorithm, Storm Relative Mean Radial Velocity task, and the Combined Attributes Table, which,
in turn, passes the storm attribute information to alert processing and the Radar Coded Message.
Some algorithms have outputs that are written to an intermediate file and are not currently used in
production of a product.
3.3.4.3.3 Products Produced from the Outputs of Other Products. As noted in

Section 3.3.4, there are products that are composed of the outputs from other products. These
products and those that contribute to their composition are:
Product Contributing Products
Combined Attributes Table Hail, Storm Structure,

M, MD, TVS, STI
Radar Coded Message Combined Attributes Table

CR Polar Grid
User Alert Message Combined Attributes Table

CR Cartesian Grid
One-Hour Precipitation Accumulation
Severe Weather Probability
Velocity
VIL/Echo Tops.
3.3.5 Product Distribution. After being generated, products are stored temporarily, on
disk, in a product database. The distribution function reads the products from the product database
to satisfy user requests. All products are required to be stored for at least two volume scans in the
event product transmission delays prevent some products from being distributed immediately after
generation. Furthermore, the database is sized large enough to accommodate storing products for
up to 6 hours to satisfy user one-time product requests.
3.3.5.1 Temporary Storage and Product Storage Loadshedding. After a

product is generated, the product is stored within a product database. The amount of time the
product resides in the database depends on the storage time specified in the RPG Product
Generation Table and whether or not Product Storage Loadshedding occurs. If no storage time is
specified, the product will be stored for a minimum of two volume scans. If a storage time is
specified for a product type, the minimum time can be as short as 30 minutes or as long as 360
minutes. See Figure 3-3 for the types of RPG loadshedding.

3-26
The product database is sized to hold at least 6 hours of products in a full load scenario. When a
product is generated, the product is marked with an expiration time which is the sum of the product
generation time and storage time. When a product has been stored beyond its expiration time, the
product is marked for expiration. For Product Storage Loadshedding purposes, the current
utilization is the ratio of the number of products in the database not marked for expiration to the
total number of products the database can hold. Like with all loadshed categories, there is a
loadshed Warning and Alarm Level (specified as a percent) for Product Storage. If the number of
products in the database that are not marked for expiration exceeds the Alarm Level, then a
sufficient number of products are marked for expiration in order to reduce the current utilization
down to the Warning Level. This event triggers the Product Storage Loadshed Alarm.
When a product is marked for expiration, is it still available in the product database. Only when
the product database is full (the total number of products in the database, either marked for
expiration or not, reaches the maximum), is an older product removed to make room for the newer
product. The product that is removed is one that has been marked for expiration.
3.3.5.2 Product Distribution and Product Distribution Load Shedding. After a

product is generated, the Product Distribution module is notified that the product is ready for
distribution. This module coordinates the distribution of all data (products, messages, and alerts) to
External User and display systems.
After being notified that a product has been generated, the Product Distribution module examines
each External User request list. For each list containing the product, the Product Distribution
module places information about the product in the distribution queue corresponding to the
External User. The position in the queue is based on transmission priority and the product priority
defined in the Product Attribute Table. The transmission priorities (highest to lowest) are defined
as follows:
5. Alert Messages
4. Product Request Response and General Status Messages
3. Alert products and User Alert Messages
2. One-time products and Routine products (High Priority)
1. Routine products (Low Priority)
Product priority is a number between 1 and 100 and varies from product to product. The value
assigned to a particular product depends on the perceived value of the product to the user
community, with higher numbers denoting higher priority or value products.
A product queued for transmission on a communications line is transmitted based on its position in
the distribution queue. That is, higher priority products are transmitted prior to lower priority
products. Consequently, lower priority products could conceivably wait in the distribution queue
long periods of time on heavily utilized communications lines.

3-27
Distribution load shedding is based on length of time in the distribution queue for a particular
communication line and prevents long transmission queue wait times. When a new product enters
the distribution queue for a communication line, a check is made to see whether any product has
been sitting in the queue for a time exceeding the product “Alarm_level * volume_time.”
If true, then products are shed until there are no products that have been waiting in the distribution
queue for a time exceeding the product “Warning_level * volume_time.”
Distribution load shedding attempts to remove products from distribution depending on age or
length of time in the distribution queue with older products (and, consequently, those of lower
priority) shed first.

3-28
CHAPTER 4
OPERATIONAL APPLICATIONS
4.1 Introduction. This chapter discusses potential operational uses of the WSR-88D products as
well as the structure of certain meteorological phenomena. It provides information on observing
and forecasting meteorological phenomena using WSR-88D products. Also included in the early
sections is a discussion of operationally troublesome radar returns due to such things as ground
clutter and sidelobes. Pictorial representations of some applications have been included.
Basic meteorological skills should be applied to determine the best products to use under different
meteorological situations. The phenomena discussed within this chapter are not all-inclusive and
by no means is the forecaster limited to products talked about in certain situations in the chapter.
As more information continues to become available it will be incorporated in or result in changes
to this chapter.
While the obvious emphasis of this Handbook is the WSR-88D application and use, it is important
to recognize that the WSR-88D is but one tool available to the meteorological practitioner or
scientist. Other tools and data sets include, but are not limited to, atmospheric soundings, surface
and upper air conventional observations and their derived fields, numerical models, satellites,
lightning detection systems, and visual observations. In order to gain the most complete
understanding of the state and evolution of the atmosphere, a synthesis of all observations is
necessary. Thus, while not specifically stated in the following sections, it is assumed that the
reader and radar user recognizes this fundamental concept.
4.2 Non-Meteorological Radar Echoes. This section briefly describes methods of identifying and
assessing the impacts of ground clutter, anomalous propagation (AP), sidelobes, and sunrise and
sunset effects. In addition, recognition methods for and impacts of range folded and incorrectly
dealiased velocity data are discussed.
4.2.1 Ground Clutter. Prior to calculation of reflectivity, velocity, or spectrum width,

return signals from ranges within the radar's normal ground clutter pattern are processed to remove
most of the signal returned from targets that are stationary (Part B of this Handbook). The clutter
signal not removed, called residual clutter or clutter residue, will remain as part of many of the
products. Normally, the clutter residue echoes are weak, in the range of 10 dBZ to 30 dBZ.
4.2.1.1 Recognition of Residual Ground Clutter. A low elevation Reflectivity

product will show ground clutter close to the radar or distant mountainous terrain if intercepted by
the beam. It will normally appear as a cluster of points (having a speckled nature) or as a large area
of contiguous returns (Figure 4-1). Residual clutter is often characterized by relatively weak
echoes, random velocities, and random but usually rather broad spectrum widths. A time lapse of
Reflectivity products will show no movement of these returns. With small, but increasing antenna
elevations, ground clutter returns will very often rapidly disappear.

4-1
Figure 4-1
Residual Ground Clutter
This WSR-88D Reflectivity product (NCDC NEXRAD Viewer) shows "residual" ground
clutter in excess of ~30 dBZ remaining, even after filtering all bins using high suppression.
Range rings are at 25 and 50 km. Radials are at every 45° of azimuth.

4-2
4.2.1.2 Assessing Impacts of Residual Ground Clutter. Residual ground clutter
near the radar may be recognized by its speckled appearance. When it is imbedded in a weak
meteorological signal over the radar and as a result of the random velocities in the clutter residue,
the Velocity Dealiasing Algorithm may introduce errors in the velocity field due to large radial and
azmuthal gate-to-gate shears greater than the Nyquist velocity. This problem is most likely to
occur in the Clear Air Mode using VCP 31 where the Nyquist velocity is about 10.8 m s-1 (21 kts)
and the gate length with the long pulse is 0.75 km (0.4 nm). The net affect of long pulse is in an
apparent increase in radial gate-to-gate shear amplitude and increase in dealiasing failures.
Although less frequent, if this problem occurs in the Precipitation Mode, the induced shears may be
picked up by the legacy Mesocyclone, Mesocyclone Detection Algorithm (MDA), or the TDA, and
carried as a feature.
Clutter residue can also have deleterious affects on other products, as well. For example, in the
Precipitation Mode, the Precipitation Processing algorithms will interpret these echoes as
precipitation and accumulate, over time, large amounts of rainfall where none has fallen.
4.2.2 Ground Clutter Returns from Anomalous Propagation. Anomalous propagation

of the radar beam is caused by non-standard atmospheric temperature or moisture gradients (Part B
of this Handbook). Superrefraction, which is frequently caused by temperature inversions, bends
the beam toward the earth and can cause the radar to detect ground returns from distances far
exceeding the normal ground clutter area. Anomalous propagation surrounding the radars is
especially common in the desert southwest and develops shortly after sunset and during the
nighttime hours because of the rapid cooling of the dry boundary layer and development of a sharp
radiational temperature inversion.
4.2.2.1 Recognition of Ground Clutter Returns from Anomalous Propagation.

Anomalous propagation is characterized by many very small “popcorn echoes” with very high
reflectivity (abnormally high for the small echo size) and extreme reflectivity gradients (Figure 4-
2). Even with small changes but increasing antenna elevation, these returns will usually rapidly
disappear. A time lapse of Reflectivity products may show rapidly changing patterns but without
echo motion. In the absence of precipitation there can be a 20 dBZe, or more, difference between
adjacent returns, mean velocities near zero, and narrow spectrum widths. Ground returns from
anomalous propagation mixed with precipitation may result in large spectrum width values and low
velocities. Erratic movement of ground returns from AP, in comparison with the motion of
precipitation echoes, may also be seen.

4-3
Figure 4-2
Anomalous Propagation
Norman, OK WSR-88D (KCRI, test bed radar) at 11:45 UTC on 6 June 2005 (CODEview
graphics). The major areas of Anomalous Propagation (AP) are induced by a nocturnal
inversion.

4-4
4.2.2.2 Assessing Impacts of Ground Clutter Returns from Anomalous
Propagation. Ground returns from AP primarily impacts precipitation accumulation algorithms
wherever the AP is located. But super refraction of the radar beam also frequently occurs behind
the cold air outflow regions of thunderstorms. As in the case of clutter residue, in these instances,
the Precipitation Processing algorithms will often erroneously interpret the ground returns as
precipitation echoes and overestimate the precipitation with these storms. It may also affect other
algorithmic output, e.g., if reflectivity is large, erroneous identification of a storm may occur
(but is rare) and its centroid location (and therefore motion) will be altered. Additionally, it can
be more difficult to interpret reflectivity echoes in the affected areas.
4.2.2.3 Removal of Ground Clutter Returns from Anomalous Propagation.

The Radar Echo Classifier (REC) algorithm has been introduced to remove or flag AP (see Part C
of this Handbook). While the REC assigns a likelihood estimate for echo classification, it offers a
means of identifying AP in real-time. Additionally, the low-level layer of the Layer Composite
Reflectivity product can be selected such that it omits the echoes from the lowest 1.52 km (5000
ft). This is also intended to remove residual ground clutter and AP. Finally, AP can be removed,
to a large extent, by application of the clutter filter to the elevation angles affected. This is
accomplished by overriding the clutter map resident in the RDA through editing of the Clutter
Suppression Regions menu at the MSCF. Up to 15 clutter suppression regions can be edited in
which three levels of suppression can be defined for the reflectivity and Doppler channels. Unit
Radar Committee agreement on the use of the clutter map editor must be developed.
When weather is not a factor, i.e., operating in the Clear Air Mode or ground clutter or anomalous
propagation are in a precipitation-free sector in the Precipitation Mode, it is reasonable to apply the
clutter filter. It is also possible to apply the filter even in precipitation regions. In most conditions
the clutter filter should be applied.
4.2.3 Sidelobes. An occasional source of data contamination is simultaneous reception of

signals at comparable power levels through both the main antenna pattern and its sidelobes (Part B
of this Handbook).
4.2.3.1 Recognition of Sidelobes. While obvious WSR-88D sidelobe returns are

rare, they can be found to the right, left, above, and below high reflectivity areas. Potential
interference from sidelobes can be diagnosed by knowing the power difference between the main
beam and the sidelobe. The WSR-88D one-way sidelobe is nominally 29 dB down. The location
of potential sidelobe interference will be slightly offset from the axis of the main beam (Part B of
this Handbook). In most cases, the velocity field of sidelobes will display noisy or erratic values.
Spectrum widths will often be extreme values and are often the best indicator of sidelobe
interference. Typically, two-way reflectivity differences between the mainlobe and the sidelobe
will be 50 to 54 dBZ for sidelobe detection to be a problem. Thus, owing to beamwidth
considerations, strongly reflective storms with very strong reflectivity gradients must be within
moderate ranges of the RDA for sidelobes to be a problem.

4-5
4.2.3.2 Assessing Impacts of Sidelobes. The presence of sidelobes may
indicate erroneously high storm tops or new storm growth where there is none. Sidelobes can
also impact velocities in weak echo by providing noisy or erroneous values that mask true
velocity patterns. Algorithmic output, especially of the TDA and MDA, may be affected in
rare cases.
4.2.4 Solar Effects. Due to the sensitivity of the WSR-88D, anomalous returns near
sunrise or sunset usually appear along one to two radials as the radar scans the sun’s location.
These returns are generated because the sun radiates energy in the same microwave region of
the electromagnetic spectrum that is used by the WSR-88D.
4.2.4.1 Recognition of Solar Effects. When the antenna scans the sun, a long
radial of weak echo return is displayed. These returns are most often observed at both local
sunrise and sunset. These echoes may be expected to appear as continuous returns, in a narrow
"baseball bat" shape, out one or two radials, at the solar altitude. Reflectivity values generally
range between 5 and 20 dBZe (Figure 4-3) and will typically increase with range. Solar effects
will appear for one or two volume scans at a single elevation of the base products and, in very
rare situations, for up to 30 minutes on a Composite Reflectivity (CR) product.
4.2.4.2 Assessing Impacts of Solar Effects. In the absence of other echoes,

typical reflectivity values from solar effects are from near zero dBZe in close to the radar to 20
dBZe or higher at 460 km (248 nm). The apparent reduction in the sun's signal nearer the radar
is due to the range normalization correction applied to reflectivity. The velocity and spectrum
width fields indicate range-aliased data out to the maximum range of these products, i.e., 230
km (124 nm). Another reflectivity-derived product that will show this contamination is the CR
and, only rarely, the Layer Composite Reflectivity or Echo Tops products. Most
meteorological algorithms will not be affected, since the reflectivity values are below the
significant thresholds and the occurrence is confined to one or two volume scans and one or
two azimuths.
Velocity-derived products that will be contaminated are the SRR and SRM. In very rare
instances and at longer ranges, the contaminated radial may disrupt pattern vectors used to
identify circulations in the MDA and TDA algorithm, where the sun's signal is stronger than
the corresponding storm echo. At close-in ranges, the sun's signal is too weak to impact most
algorithms.

4-6
Figure 4-3
Solar Effects
Mosaic of WSR-88D Composite Reflectivity from 12:34 UTC on 21 April 2003. The streaks
of low reflectivity extending outward to the east northeast from the radar locations are
caused by sunrise effects. That is, the radar is detecting radiation incident from the sun in
the same wavelengths as the radar backscattered power. Source: NWS Meteorological
Development Laboratory.

4-7
4.2.5 Range Folded Data. Range ambiguities in the Doppler velocity and spectrum
width fields caused by the WSR-88D's pulsed Doppler sampling interval (Part B of this
Handbook) can be very significant. For any pulsed Doppler system, the product of the
unambiguous range and the Doppler Nyquist co-interval is a constant function of the
wavelength of the radar and the speed of light. By decreasing the PRF, the unambiguous range
of the radar can be increased; however, the Nyquist velocity range interval will decrease. The
WSR-88D's unambiguous range for Doppler velocity and spectrum width measurements varies
from 120 km (65 nm) (at a Nyquist co-interval of + 31 ms-1 (60 kts)) to 174 km (94 nm) (at a
Nyquist co-interval of + 21 ms-1 (41 knots)). The reasons for range folding and methods to
dealias the velocities are covered in Part B of this handbook. However, it can be summarized
here.
The WSR-88D scanning strategies combine Doppler waveforms having high PRF and short
unambiguous ranges with a surveillance cut with low PRF and long unambiguous range in order to
attempt to resolve range ambiguities of the Doppler scans. The surveillance waveform is
unambiguous out to a 248 nm range. One dealiasing method uses the “split-cut,” which first scans
an elevation with the surveillance waveform and then scans it a second time using the Doppler cut.
The second method, the “batch cut,” interleaves the surveillance and Doppler waveforms. In
each case the dealiasing algorithm uses the surveillance waveform to do a gate by gate power
comparison of the Doppler waveform multiple “trip” ambiguous ranges. This gate by gate power
comparison will reveal which, if any, of the ambiguous range gates dominates. In cases where
there is potential range ambiguity or overlaid echo, to summarize the algorithm processing, those
gates that dominate in power returned are those range gates determined to be where the Doppler
information is originating. The remaining gates with significantly weaker power returns are
displayed as “overlaid data” and unusable for Doppler measurements. If returns are present, but
none of the ambiguous gates dominate in power, then the data is determined to be range overlaid
for all those gates.
In more detail, the dealiasing algorithm functions in the following manner. The radar interprets the
locations of velocity and spectrum width returns from beyond the unambiguous range as occurring
within that range, as well as beyond. The range-unfolding algorithm has been implemented in the
RDA preprocessing to attempt to replace range-folded Doppler data from as many as four trips to
their proper locations in range. The range-unfolding algorithm compares the return power from all
possible ranges given the Doppler PRF used for that slice. This is done range gate by range gate
for all possible ambiguities. If the power return from one possible range gate exceeds the power
return from all other possible ambiguous or “matching” range gate locations by more than the
defined threshold (TOVER, default 5dB), the Doppler data are assigned to that range and all other
possible ranges are flagged as range obscured (overlaid). If the power return from one possible
range does not exceed the power return from the other possible ranges by the defined threshold
(TOVER), all ranges are considered range obscured (overlaid) and the Doppler data are not
assigned to any of the possible ranges. Range obscured (overlaid) data are flagged as range folded,
treated as missing by all downstream algorithms, and are displayed as purple (adaptable) in the
display (Figure 4-4).

4-8
Figure 4-4
Range Overlaid Echoes
Dodge City, KS WSR-88D Base Radial Velocity Data Array product at 03:00 UTC on 30
May 2004 (AWIPS display). This product depicts significant range obscured (overlaid) or
range folded echoes (color coded purple), mostly northeast through south of the radar. A
strong cold front and associated wind shift is passing the radar.

4-9
4.2.5.1 Recognition of Range Folded Data. Within the WSR-88D, range folding
affects only the Doppler data, that is the velocity, and velocity spectrum width data and products.
However, through the use of the surveillance scan, the affected Doppler data are flagged and color-
coded.
Range overlaid echo is especially a problem when the near-radar ground clutter residue or even
clear air echo is repeated just beyond the edge of the 2nd and 3rd trip echoes. This can account for
substantial range folding in each range interval. Range folding may also occur under conditions of
anomalous propagation where the radar beam is constrained to follow a path close to the Earth's
surface. Range folding is also a significant problem when strong convection occurs beyond the
first trip or within widespread precipitation echo such as with tropical storms.
When possible, the range-unfolding algorithm will place the Doppler velocity and spectrum width
data at the proper range. When this software cannot determine the proper range, i.e., the data is
determined to be overlaid, the data will be flagged and displayed as range folded.
An alternative dealiasing method employs the MPDA (Part C, Chapter 3, of this Handbook),
currently confined to VCP 121. Its primary application is velocity dealiasing, but it is also
effective in range unfolding. The Doppler data for the first PRF are automatically range unfolded
by the RDA. Doppler data for the 2nd and 3rd scans must be range unfolded at the RPG using the
MPDA preprocessor algorithm. This algorithm reduces range folding by 50% to 70%. The
algorithm implements three different PRFs and monitors the changes, if any, of Doppler data
placement. In this way, the correct data placement can be determined.
4.2.5.2 Assessing Impacts of Range Folded Data. Range folded data can impact
products and algorithms that use velocity data. However, there is essentially no impact on
reflectivity data and products.
The impact of Doppler velocity and spectrum width range folding is significant, both in the
Doppler waveform’s unambiguous range limits and in 2nd or 3rd trip areas where significant
velocity data loss due to range obscuration (overlaid) returns also occur. Range obscured
(overlaid) data are flagged as range folded and treated as missing by the velocity-based algorithms
and can, therefore, seriously impact those algorithms. But perhaps more importantly, overlaid data
significantly impacts the human interpretation of the data.
4.2.5.3 Assessing Impacts of Range Folded Data on Velocity Products. The

presence of range folded (overlaid) data on Mean Radial Velocity products is inevitable. The
inability to determine velocity estimates for these sample volumes results from the inability of the
range unfolding algorithm to distinguish between nearly equal power returns from two or more
sample volumes at the same relative location within different trips. Therefore, valid velocity
estimates can be derived for only one corresponding sample volume along each radial if the power
returned from the sample volume dominates. If the power returned from any of the ambiguous
sample volumes fails to dominate, then all will be flagged and color-coded as range overlaid echo.

4-10
The ring of range folded (overlaid) data at the beginning of the second and subsequent trips is
caused by the first trip ground clutter and is a common result of this range unfolding limitation.
The result, as mentioned above, is the inability of either the human or the algorithm to interpret or
use this overlaid velocity or spectrum width data.
4.2.6 Velocity Aliased Data. Velocity aliasing occurs when frequencies too high to be
analyzed with the given sampling interval appear as a frequency less than the Nyquist co-interval.
In other words, wind speeds greater than the unambiguous velocity (Nyquist co-interval) for the
current PRF are wrapped around into the incorrect Nyquist co-interval. A sophisticated velocity
dealiasing technique is implemented in the WSR-88D (Part C of this Handbook). However,
improperly dealiased data (Figure 4-5) occasionally occurs where the first sample volume
intercepted by the radar beam already exceeds the unambiguous velocity or where gate-to-gate
shears are very large. What we deal with is not aliased data, but with incorrectly dealiased data or
dealiasing failures.
4.2.6.1 Recognition of Velocity Dealiasing Failure. Data that are incorrectly

dealiased can often be identified by inspection. Typically aliased velocities that, for some
reason, cannot be dealiased properly are actually passed through as incorrectly dealiased.
Velocities may appear reasonable until the sample volume is encountered where the failure occurs.
Errors are recognizable as wedges or radials having sharp radial discontinuities from adjacent
regions and whose predominant colors differ markedly (Figure 4-5). Errors may also appear as
radial spikes several volume samples in length, whose velocities are shifted toward high positive or
negative values. When it is noted that there is no zero crossing between these high values of
opposite sign along a radial or between this and the adjacent radial, then it is certain that a
dealiasing failure has occurred. Moreover, if the maximum listed velocities in a velocity product
are shown as +/- 62.8 ms-1 (122 kts) or +/- 63.2 ms-1 (123 kts), this often (but not always) signals
that incorrectly dealiased velocities are present somewhere in the product.
Instances of incorrect dealiasing may occur when there are shifts in the inward and outward bound
velocities along the radials of data that do not fit those allowed by the algorithm. In these cases,
the actual values may be off by a factor of twice the unambiguous velocity of the PRF in use at that
time. Typical unambiguous velocities for the WSR-88D, in the Precipitation Mode, range from
20.6 ms-1 (40 kts) to 30.9 ms-1 (60 kts). Very rarely, groupings of data appear along a small set of
radials that could not be successfully dealiased. These should be obvious. However, the vast
majority of time, velocity data having passed through the dealiasing algorithm is altered to the
point that the non-dealiased data and its actual value are irrevocably lost.

4-11
Figure 4-5
Velocity Dealiasing Errors
Fort Worth, TX WSR-88D 4-panel, Base Velocity Data Array product at 00:07 UTC on 6
April 2003 (AWIPS display). There are extensive areas of velocity dealiasing errors
southwest of the sharp velocity discontinuities from mostly moderate radar inbound (green)
velocities adjacent to the very strong (yellow and pink) outbound velocities. There is no
zero crossing along these radial-aligned velocity discontinuities giving indication that there
is a dealiasing error in that location.

4-12
4.2.6.2 Assessing Impacts of Velocity Dealiasing Failures. Incorrectly dealiased
velocity data can seriously impact certain WSR-88D algorithms and products. Velocity products
of all types (V, SRM, SRR, VCS, SWV, DV, and ITWSDBV) will be difficult or impossible to
analyze when contaminated with incorrectly dealiased Doppler data. This will significantly
degrade the front line of human data interpretation. In order to determine the extent of the
dealiasing problem, it is recommended that earlier displays of these products be examined to
determine if there is temporal or spatial continuity. In addition, other elevation angles of the
velocity products may be used to determine if there is vertical continuity.
Algorithms and products that ingest mean radial velocity data can output incorrect results when
such data are used. In the case of the legacy Mesocyclone, MDA, and TDA algorithms, there will
likely be a lack of vertical continuity of incorrectly dealiased data. Consequently, only
uncorrelated shears should result from using incorrectly dealiased velocity data. In the rare event
of contaminated Mesocyclone, MDA, or TDA output, other products should help verify the
existence of these circulations. However, base products should always be the primary tool for
mesocyclone and Tornado Vortex Signature (TVS) detection in any case. (Algorithm output
should function only as an imperfect “safety net”).
4.3 Non-Precipitation Radar Echoes. The WSR-88D provides the capability to monitor the
atmosphere in the absence of precipitation-size backscattering particles in the optically clear air.
Clear air return is very useful because it will often allow measurement of low-level environmental
winds (i.e., those within the boundary layer) in all but the coldest months in moderate or cold
climates. The capability can also prove important in recognizing early or pre-convective
development, turbulence, and the onset of severe wind events.
These “clear air” returns are associated typically with two target types. The first source of clear air
backscattered power is the refractive index variation or density gradients within the atmosphere.
These are caused by temperature and humidity variations. Reflectivity with these variations is less
than about 5 dBZ. The second source is associated with insects, birds, and sometimes bats (Doviak
and Zrnic, 1984). Reflectivity often varies from 5 dBZ up to over 50 dBZ depending on the size
and concentration of these biological targets. However, there is also smoke, dust (with sufficient
particle size for S-band radar detection), volcanic ash, chaff, and other sources of non-
meteorological echo (Figure 4-6).

4-13
Figure 4-6
Dust Storm Reflectivity
Lubbock, TX WSR-88D Base Reflectivity Data Array product at 19:35 UTC on 19 February
2004 (AWIPS display). A dust storm is responsible (with significant refractivity gradients
and sufficiently large particulates for S-band detection) for this large area of Clear Air Mode
echo. The overlaid surface observations (20:00 UTC) indicate wind gusts ranging from
~15 to 25 ms-1 (30 to 50 kts), blowing dust, and sky obscuration due to blowing dust.

4-14
4.3.1 Recognition of Non-Precipitation Radar Echoes. Non-precipitation echoes can be
seen in the Precipitation Mode as well as the Clear Air Mode. In the Clear Air Mode non-
precipitating returns will be seen out to about 185 km (100 nm) in extreme cases, depending on
environmental conditions. In the Precipitation Mode, it may be expected that recognition of non-
precipitation echoes will be confined to ranges closer to the antenna (out to ~ 46 km (25 nm) to
extremes of ~ 130 km (70 nm)). In most radar locations insects and birds increase markedly at
sunset and the convective boundary layer is seen to “bloom” on radar. At other times migratory
birds can be detected and even tracked by radar (Kelly et al. 2000; Bruderer 1977a, 1977b, 2000).
Using the WSR-88D, it is possible to attain significant radar measurements in optically clear air.
With enough atmospheric scattering, the WSR-88D can measure a return of -8 dBZe at a range of
50 km (27 nm) and -28 dBZe at a range of 5 km (2.7 nm). When operating in the Clear Air Mode,
use can be made of the 16-level R and V products, as well as SW and VAD products.
4.3.2 Considerations. In the Clear Air Mode, the option exists to select either long or
short pulse (VCP 31 or 32, respectively). Operation in the long pulse results in a better signal-to-
noise ratio permitting lower reflectivity levels to be detected (-16 dBZe at 50 km (27 nm)). The
long pulse, while providing better sensitivity, results in coarser resolution in velocity and spectrum
width fields (~0.75 km (0.4 nm) range gate resolution) and a lower Nyquist velocity value (+ 10.8
ms-1 (21 kts)). In Clear Air long pulse Mode (VCP 31), the actual pulse width is 0.75 km
(0.4 nm) which restricts velocity and spectrum width product resolutions to 1 km (0.54 nm).
The long pulse mode is best suited to situations where the atmospheric boundary layer is well
mixed, such as during daytime heating. In the short pulse, the sensitivity is only slightly reduced
and the ~250 m (0.13 nm) gate resolution is maintained in the velocity and spectrum width
fields. Since the higher Nyquist velocities are available in the short pulse, this VCP may be
used when strong winds characterize the environment.
4.4 Boundaries. Various airmass or airflow boundaries can readily be seen on the WSR-88D.
Some of these boundaries include: synoptic fronts, gust fronts, pre-frontal trough lines, drylines,
thunderstorm outflows, land and sea breezes, and other convergence zones that may be inherent to
particular regions. While the mechanisms for inducing each of these boundaries may be somewhat
varied, their presentation on the displays are similar enough to discuss them in a broad sense
collectively. Examples can be seen in Figures 4-7 through 4-9.
4.4.1 Recognition of Boundaries. In general, boundaries are characterized by narrow

zones of shifting convergent winds or density differences, or both. They will appear on
Reflectivity (Figure 4-7) and Spectrum Width products (Figure 4-59) as narrow lines of enhanced
values (sometimes called thin lines). The enhanced reflectivity is produced by a combination of
converging particulates, e.g., insects and the birds and sometimes refractive index gradients from
temperature or moisture differences or both. Enhanced spectrum widths are typically produced by
wind shear, changes in wind speed and direction within individual sample volumes.

4-15
Figure 4-7
Dryline Reflectivity
Amarillo, TX WSR-88D Reflectivity product at 00:00 UTC on 6 May 1993 (legacy PUP
display). This product, collected in Precipitation Mode, shows a linear thin-line echo along
the “dryline” just to the west of the radar and convective storms to the north associated with
the boundary. The thin-line echo is likely associated particulates, insects and birds, and
with a change in refractivity associated with the dryline.

4-16
Figure 4-8
Boundaries and Severe Convection Reflectivities
Hastings, NE WSR-88D Base Reflectivity Data Array product at 00:26 UTC on 23 June
2003 (AWIPS display). This product, collected in Precipitation Mode, shows two
boundaries. The thin-line to the left of the radar is along a stationary frontal boundary while
that to the right is along an outflow boundary from earlier thunderstorm activity. The
boundaries are delineated by echoes likely associated with insects, birds, and refractivity
differences. The storm complex to the north northeast of the radar had, shortly before this
image, produced the largest hail ever recorded, measuring 7.0 inches (17.78 cm) in
diameter with an 18.75 inch (47.63 cm) circumference, and weighed 1.33 pounds (0.6 kg).
The complex to the southeast of the radar produced the strongest mesocyclone ever
recorded (Wakimoto et al. 2004).

4-17
Figure 4-9
Dryline – Cold Front Intersection Reflectivity
Oklahoma City, OK WSR-88D Reflectivity product at 22:11 UTC on 24 April 1993 (legacy
PUP display). This product, collected in Clear Air Mode, shows thin-lines associated with a
cold front (northwest of the radar), dryline (east of the radar), and a third boundary-like
feature of unknown origin between the other two. Echo intensities less than about 5 dBZ
are associated with refractive index gradients. Convective echoes soon developed near
the “triple point” of boundary confluence and led to a tornadic supercell storm.

4-18
The changing winds across the boundary will show up on Mean Radial Velocity products as radial
convergence or azimuthal shear, depending on the orientation of the winds with respect to the
radar-viewing angle. When the radar beam is perpendicular to the long axis of the boundary, radial
convergence will be detected. When the beam is parallel to the boundary, the wind shift will
produce a long, narrow zone of azimuthal shear and broadened velocity spectrum width.
Boundaries of sufficient horizontal extent will be defined by a combination of convergence and
shear on the same display.
4.4.2 Considerations. Boundaries will be better detected in the Clear Air Mode as
opposed to the Precipitation Mode because of increased sensitivity with clear air sampling
parameters. This is especially true with the increased sensitivity of the long pulse employed in
VCP 31. At 50 km (27 nm) range, returns of -16 dBZe can be detected in the Clear Air Mode, but
when in Precipitation Mode the lower bound reflectivity displays are typically limited to 5 dBZe
and higher using the 16-level Reflectivity product (Figure 4-7). However, using the DR product,
the lower bound reflectivity threshold, even in the Precipitation Mode is -30 dBZe and is clearly
applicable for boundary detection (Figure 4-8). In the Clear Air Mode, boundaries are best
detected by the 16-level Reflectivity product which, in this mode, as indicated above, has a much
lower threshold than the same product in the Precipitation Mode (Figure 4-9). In the
Precipitation Mode, boundaries can often be detected, but not as well as in the Clear Air Mode;
boundaries away from precipitation areas will most frequently be seen closer to the radar (often
within ~ 55 km (30 nm) or less) unless using the DR product. Gust fronts in association with squall
lines and fronts that are attended by mesoscale precipitation bands most often have reflectivity
factors of +5 to +25 dBZe and their boundaries will sometimes be detectable at longer range
increments.
Display system software that implement “feature tracking” can be used to calculate boundary
motion. Time lapse loops can be used for this purpose, as well. In addition, the intersection or
collision of boundaries, which may lead to further convective development, can often be easily
anticipated or observed by viewing time lapse sequences.
When a boundary passes the radar and a frontal “airmass” covers the radar for some distance, then
the VWP product may be used to detect the cold air depth (typically 1 to 3 km (3,000 to 10,000 feet
deep)), its changes, and vertical frontal structure. The vertical extent of land or sea breezes can
also sometimes be determined using the VWP as an indicator.
4.5 Cloud Layers. Strictly speaking the WSR-88D, because it is an S-Band weather radar, cannot
detect cloud droplets owing their very small size. This is the case with warm clouds that is clouds
above freezing that do not contain precipitation. However, often and under certain circumstances
the sensitivity of the WSR-88D provides the user with the capability of detecting cloud layers due
to precipitation size particulates (~0.1 to 1 mm) or refractive index gradients. The WSR-88D has
revealed that many clouds, even some which are translucent, contain S-band radar detectable
precipitation-size particles. Moreover, refractive index gradients sufficient to create radar returns

4-19
can reveal a variety of other clouds. Smoke plumes and smoke layers (with sufficiently large
particulates) are often revealed in the same way as are volcanic ash plumes. The Reflectivity
products in both Clear Air and Precipitation Modes are commonly used to determine the existence
and extent of the cloud layers. Thus, in this Handbook when referring to WSR-88D “cloud”
detection, we are considering detection of either precipitation-size particles within cloud or cloud-
associated refractive index gradients. This capability is important because it has application to
aviation forecasting, and forecasting or observing the early evolution of precipitation.
The WSR-88D can detect large ice crystals and precipitation size particles that are present in
middle and high-level clouds. Of course, virga descending from these same clouds is also
detectable. Reflectivity may range up to and in excess of +30 dBZe within these precipitation
bearing clouds. Thunderstorm anvil “blow-off” and middle- and upper-level precipitating cloud
layers associated with stratiform precipitation and warm fronts are prominent examples. Because
these clouds can be easily detected, other aspects of the middle troposphere can be observed. The
VAD algorithm, for example, will use Doppler velocity measurements from middle and high-level
clouds to generate profiles into the middle troposphere. These wind profiles allow the operator /
meteorologist to observe the movement of synoptic and smaller scale waves and troughs and can be
used to monitor numerical model performance.
4.5.1 Recognition of Cloud Layers. With the high sensitivity of the WSR-88D, it is
possible to obtain reflectivity estimates within clouds at all heights which generally reflect at levels
between -12 and +5 dBZe (Figure 4-10). Moreover, it has been observed that refractive index
gradients within atmospheric layers that are undergoing cooling and lifting prior to actual cloud
formation are also sometimes detectable.
The depth of cold cloud layers may often be inferred from particulate detection and warm cloud
layers from detecting refractive index gradients. Cloud layers very often appear as concentric rings
about the radar location. The inner edge of the ring marks the cloud base while the outer edge
marks the cloud-layer top. Using the Reflectivity product for the elevation that intersects the cloud
layer, place the cursor at the point where the base appears to be and note the azimuth, range, and
height. Then, place the cursor at the apparent top of the cloud layer noting the same information.
This reveals the cloud depth at those points. These layers are, however, often not of uniform depth
and may slope with height (Lemon and Quoetone 1994).
4.5.2 Considerations. In the Clear Air Mode, cloud detection may be best achieved using
VCP 31 due to a better signal-to-noise ratio. As explained above, cloud droplets characterizing
warm clouds (stratus and fog absent of precipitation) are too small to be detected by S-band radars.
Horizontal cloud rolls or cumulus cloud streets can also occasionally be detected with reflectivity
often ranging from -10 dBZe to +10 dBZe. This is again because of refractive index gradients or
convergence of particulates such as insects, birds, etc. Gravity waves and undular bores atop an
inversion can also be detected at times in the velocity and reflectivity fields. Gravity waves should
coincide with parallel bands of alternating increases and decreases in reflectivity.

4-20
FEBRUARY 2006
4-21
Figure 4-10
Clear Air Reflectivity Product
FMH-11-PART-D
Oklahoma City, OK WSR-88D Reflectivity product at 17:13 UTC on 12 December 2000 (NCDC NEXRAD viewer). In this
product, collected in Clear Air mode, the radar-centered “donut” echo is an ice crystal cloud and a virga layer aloft surrounds
the radar.
4.6 Pre-Convective Development. Of importance to all forecast programs is the ability to observe
and forecast development of clouds and convection.
4.6.1 Recognition of Pre-Convective Development. In areas of warm advection, small,

weak, echoes may begin to appear in bands or clusters. These echoes will then increase in numbers
and intensity at various rates. In other cases lines of broken convection become apparent when
these small, weakly reflective echoes rapidly increase and become strong within a few volume
scans. Observations of this can be accomplished in the Clear Air Mode using all three base
moments. At other times in an otherwise noisy field, the user can begin to detect velocity
coherency and low values of spectrum width. Reflectivity returns may appear to be small and
essentially circular in nature.
4.6.2 Considerations. Evidence of convective development will sometimes appear in the

spectrum width and velocity before any significant return is detected in the reflectivity field.
Selection of the long pulse in the Clear Air Mode may produce better results because of increased
sensitivity. For the earliest evidence of convective development, observations at or near the
altitudes of the 0° C to -15° C isotherms are often used.
4.7 Convective Storms. Deep moist convection develops in a conditionally unstable environment
with sufficient moisture and lift. Lift and updraft may be produced in a variety of ways including
by warm advection, low-level convergence, cold advection aloft, jet streaks aloft, orography, etc.
Once precipitation particles form and precipitation drag and evaporation begin, corresponding
downdrafts will soon develop. It is important to realize that convection and convective storms are
compact regions of airflow and heat exchange. These storms and their corresponding radar echoes
are not “things” or objects, but rather four-dimensional airflow processes.
Many of the radar products in the WSR-88D are designed to examine convective storms and the
features that mark them as severe. The WSR-88D is a diagnostic tool in the hands of the
meteorological researcher and practitioner. Through the use of WSR-88D products, characteristics
of single cell, multicell, and supercell storms are investigated with the intent to determine storm
existence, motion, strength, and severity. In this section we discuss WSR-88D detectable storm
features and the analysis methods and products used to discriminate between severe and non-severe
storms. (The user is urged to use not only radar data, but all observations).
4.7.1 Single Cell Storms. Modern concepts of convective storms owe their origin to the
Thunderstorm Project of the 1940’s (Byers and Braham 1949). Of fundamental importance
identified in this project was the basic organizational structure of deep, moist convection that is the
“cell” (Figure 4-11). The meteorological scientific community recognizes the buoyant bubble or
plume of updraft and cumulus cloud as a thunderstorm “cell.” This rising bubble or plume gives
rise to a precipitation ensemble detectable on the radar as an echo center or echo reflectivity core.
As precipitation begins to form and descend a downdraft will quickly result through entrainment of
dry potentially cold air aloft and precipitation drag. In time, this downdraft will spread and in non-

4-22
severe convection will not coexist with the updraft for long periods but will overtake and dominate
the “cell” leading to updraft demise and demise of the cell, leaving only an “orphan anvil.”
Commonly, the life cycle of a single cell ranges from perhaps 20 to 40 minutes.
In meteorological operations, the cell concept has been somewhat modified and based on the
primary observing system, that of the radar. The thunderstorm “cell” has come to mean a discreet,
typically low-level, individual convective radar echo (see Part B, Chapter 8, of this Handbook).
Over the years, taxonomy of convective storm types has developed around this concept. Most
recently, convective storms have been seen as “ordinary” and “supercell” convection. However,
here we persist with the former system emphasizing both cell and storm organization. The first, or
fundamental, storm type in this system is the “single cell storm” (Figure 4-12). Arguably, single
cell storms do not, in reality, exist; however, for our operational purposes we will begin with what
is called the single cell storm. (We neglect for now the “new cell” development shown in Figure 4-
12.)
4.7.1.1 Recognition of a Single Cell Storm. The environment for the single cell
storm is normally one of relatively deep moisture, weak wind shear of the horizontal winds with
height, and modest instability. An ordinary, non-severe, single cell storm typically has a
radar pattern in which the reflectivity distribution is vertically aligned; no weak echo region or
echo overhang is present. Displays of successively higher elevation angles of the Reflectivity
product in a 4-panel display should confirm the vertical alignment and a narrowing of the echo
(Figure 4-13). A Reflectivity Cross Section product, generated for an appropriate axis through the
storm, may also confirm this alignment (as in Figure 4-12). Storms having this structure are not
normally severe. They may produce lightning and thunder, brief heavy rain and infrequently small,
non-severe, hail and gusty surface winds. Other products normally used are Mean Radial Velocity;
VIL; STI; HI; and CR with the Combined Attributes Table; and the OHP, THP, and the
STP/Digital Storm Total Precipitation (DSP) accumulation products.
Single cell storms generally move in the direction and with about 70% of the wind speed of the
mean wind through the cloud-bearing layer. With reflectivity values typically above 40 dBZe, and
easily identified storm reflectivity centroids or centroid, the Storm Cell Identification and Tracking
(SCIT) algorithm should adequately track movement of these storm cells. However, with the
storm’s typically abbreviated life cycle, these single-cell storms may be tracked for only 3 to 6
volume scans while the total lifetime of the overall echo will be longer.
4.7.1.2 Considerations. As single cell storms intensify, they invariably become

multicell cluster storms. Moreover, and especially when in a sheared environment, these storms
will often begin to propagate, and occasionally deviate, to the right or left of the mean flow. In this
event, the STI product will be unable to accurately project the change in its movement since a
linear track is assumed.

4-23
Figure 4-11
Ordinary Cell
The three stages of an ordinary cell life cycle, a) towering cumulous, b) mature, c)
dissipating are depicted. Features of the figure are labeled showing distributions of cloud
(outlined) encompassing updraft and precipitation encompassing downdraft. (From
Doswell 2001 after Byers and Braham 1949).

4-24
Figure 4-12
Non-Severe Thunderstorm Schematic
Schematic diagram of a non-severe thunderstorm in a sheared environment. In the upper

panel is seen the cloud outline and the radar echo and features as labeled. Placement of
the vertical cross section (A – B) in the lower panel is shown in the upper panel. Vertical
cross section includes dashed echo contours, cloud outline, and low-level inflow and high-
level outflow is indicated by the shaded arrow. After Lemon (1980).

4-25
Figure 4-13
Ordinary Non-Severe Thunderstorm
Norman, OK WSR-88D (test bed radar) 4-panel display of Reflectivity products at 01:30
UTC on 23 June 1989 (legacy PUP display). The white cross at the 0.5° elevation upper
left, 1.5° in upper right, 2.4° in lower left, and Echo Tops product in the lower right is at the
same geographical location. The white cross is at the location of the echo summit and the
high reflectivity echo core aloft of this nearly vertical thunderstorm.

4-26
4.7.2 Multicell Storms. A multicell storm consists of a cluster of single, short-lived cells
in different stages of their life cycle (Figure 4-14). Often, each multicell cluster storm is one
contiguous echo in low-levels on the display but the multicell nature of the echo can be seen in
mid- and upper-levels of the echo. The outflows from each cell frequently combine to produce a
low-level gust front, and the convergence along the boundary triggers new cell development on a
preferred storm flank. This is especially true when the storm is embedded in an environment of
deep layer shear.
4.7.2.1 Recognition of a Multicell Storm. A multicell cluster storm is evident

when two or more (often mid-level) echo cores and/or echo tops become apparent within the echo
mass (Figure 4-12). These cells will usually be in various stages of development, as indicated by
increasing and decreasing values of reflectivity and fluctuating echo tops, over time. A time lapse
of a Reflectivity product should help identify development of new cells on a preferred flank
(“updraft” flank) of the storm. As each new cell develops and matures it becomes the dominant
updraft and cell within the complex only to be replaced by successive new cells and associated
updrafts. As each cell is replaced it moves downstream as it weakens and dissipates.
In a multicell storm, an indication that there is a transition to severe activity is the appearance of a
cell with its first significant echo at higher levels than the earlier cells. However, this is often
difficult to recognize in real time.
In a sheared environment, several changes in the echo may indicate a transition to a severe
thunderstorm. Most of the echo features indicating severity will be discussed in Section 4.7.4., but
a few will be mentioned here. If the multicell storm in a sheared environment becomes severe, a
mid-level echo overhang on the updraft storm flank will develop with a weak echo region beneath.
In low-levels, a strong reflectivity gradient develops on the updraft storm flank. This overhang and
strengthened reflectivity gradient reflects invigoration of the updrafts and the subsequent
enhancement of upper level divergence. This can be recognized by displaying successively higher
elevations of the Reflectivity product in a 4-panel/multi-panel or “all tilts” display, with each panel
centered and magnified identically. This can be done in an automated fashion using Reflectivity
products on the current RPS list. Moreover, in this same fashion V, SRM, or SW can be displayed
revealing the matching kinematic storm structure. The overhang and Weak Echo Region (WER)
features can also be identified using Reflectivity Cross Section (RCS) products, but the axis and
endpoints must be very carefully chosen. In order to be certain that the three-dimensional
reflectivity storm structure has been best delineated, several RCS’s will probably be required and
even then, it cannot be certain that the proper axis was selected. This process and the serious
limitation in placement are not present with a 4-panel, multi-panel, or all tilts Reflectivity product
display. Rather, much of the structural ambiguity is removed through quasi-horizontal display of
Reflectivity multi-panel displays. Therefore, the multi-panel method of product display is
recommended over the vertical cross section method.

4-27
Figure 4-14
Multicell Storm Sequence Schematic
Schematic illustrating the evolution of convective cells moving through a multicell storm
over a 20-miute period.

4-28
As each new cell of the multicell cluster storm tends to develop in the up-shear direction, often to
the right of its predecessor, the storm as a whole will usually deviate in the direction of updraft
propagation and move more slowly than the individual cells and the mean wind. At other times
new cell development on the storm front flank may accelerate storm motion. Then again, with
downstream cell motion but upstream cell propagation offsetting each other, the result can be a
stationary storm. There are many and varied combinations of propagation and cell translation. A
time lapse of low-level Reflectivity products will help in showing deviant motion and new cell
development, often in the direction of the updraft storm flank.
When, with a multicell radar echo, the low-level echo core is displaced toward a developing
updraft-flank along with enhanced reflectivity gradients, an echo overhang, and WER aloft, with
the storm top directly above, then that storm is a developing severe storm. The dominant severe
weather phenomena are typically large hail (> 1.9 cm (¾ inch) hail) and occasionally severe or
damaging surface winds (> 25.7 ms-1 (50 kts)). Most often the severe weather occurs in an
episodic manner as each new cell developing on the updraft storm flank matures and begins to
decline.
In a squall line, a linear multicell system, the portion of the line with the most intense updrafts
can be located by using the multi-panel method to identify the area of strongest low-level
reflectivity gradient, displaced low-level echo core, the mid-level echo overhang and WER
beneath, and the displaced echo top. These features, when associated with the squall line, are
normally along the advancing edge and again indicate a severe storm. In fact, as a segment of
the line begins to intensify, the use of the multi-panel or quarter-panel display with appropriate
reflectivity slices may signal this with a shift of echo top from over a storm core to along the
leading edge of the line as the WER develops. As with other storms in a sheared environment,
these same structural features are consistent with a storm producing large hail and sometimes
damaging winds (Figure 4-15). With squall lines, however, damaging winds are a substantially
greater threat than are the very large hailstones.
Further, when squall lines or Bow Echo events are located near the radar (within about 40
miles), damaging winds may be detectable if the viewing angle is largely parallel to these
winds. At longer ranges the “Deep Convergence Zone” (DCZ) (Lemon and Parker 1996) or
the Mid Altitude Radial Convergence (MARC) (Schmocker et al. 1996) signatures can be used
to infer the presence of damaging surface winds.
4.7.2.2 Considerations. While the SCIT algorithm should readily identify the individual
storm cells, incorrect forecast movement by the STI product will occur if a storm begins to change
direction in a nonlinear fashion or, occasionally, when reflectivity components are incorrectly
assigned by the SCIT algorithm. Thus, with several cells in close proximity to each other in a
multicell storm complex, the product may not adequately track storms. When storm cell tracking
may not forecast storm movement as a whole; a time lapse of Reflectivity products will often
provide a better estimation. Correcting these limitations of the SCIT algorithm is the subject of on-
going research.

4-29
Figure 4-15
Severe Thunderstorm Schematic
As in Figure 4-12 except schematic diagram of a severe thunderstorm in a sheared

environment. After Lemon (1980).

4-30
“Deviant” storm movement as compared to the mean wind or motion of most other echoes may
indicate that the storm has become or is becoming severe. This deviant movement will best be
identified using the time lapse capability and specifying a continuous update.
As before, one method of identifying the overall echo structure is to generate a multi-panel display
of different elevation angles of Base Reflectivity centered and magnified on the storm (Figure 4-
13) or via an “all tilts” product. Reflectivity Cross Sections may also be useful, but as emphasized
earlier, placement is critical and several cross sections may need to be generated. This is both time
consuming and of questionable value owing to placement uncertainties. Another method to
determine storm structure is to build a quarter-screen display with four different elevations of
reflectivity, all centered, and magnified at the storm location. Linked cursor readout will then
enable the user to visualize the relationship of weak echo regions, overhang, and other structural
features of interest. Use of typical RPS Reflectivity products in the multi-panel display is superior
to this. In any case, use of the background maps (maps in foreground) with quarter-panel displays
and magnification will aid in WER extent recognition.
Vertically Integrated Liquid can be used for convective storm diagnostics. Rapid changes in VIL
values may signify the onset of severe activity, but gaps in the VCP elevation angles can induce
these changes strictly as an artifact of sampling. For this and other reasons, caution is
recommended in its use.
4.7.3 Supercell Storms. Supercell storms occur in unstable, highly sheared environments,
very often producing large hail, heavy rainfall, high winds, and sometimes strong or violent
tornadoes. Supercell convective storms are defined by the presence and persistence of a
distinguishing characteristic, that of the deep mesocyclone (vorticity of 1 x 10-2 s-1 or greater)
(Browning 1977). These storms are of great importance due to their disproportional production of
damage and nearly all strong or violent tornadoes. There are arguably three types of supercells
(Doswell and Burgess 1993): Classic, High Precipitation (HP), and Low Precipitation (LP)
Supercell storms. While considerable subjectivity is involved in discrimination of these types and
sampling parameters have a strong influence on radar typing, these supercell types will be
discussed here. Much of the discrimination as to supercell type deals with the location of the
updraft and downdraft, as well as the distribution, intensity, and type of precipitation near and
around the primary updraft. This, in turn, is often a function of storm relative winds over the storm
depth. For example, the so-called “Classic” supercell typically maintains updraft to the right and
rear of the precipitation cascade and downdraft which are positioned on the left storm flank. The
so-called HP Supercell typically maintains a larger and more intense precipitation cascade to the
rear of the front-flank updraft. In both cases there is the very important “Rear Flank Downdraft”
(RFD) across the storm’s rear flank and into the hook echo area. Much of the RFD in the HP case
is within the precipitation cascade region, whereas in the Classic case some of the RFD is within
the hook echo and behind; more is outside the storm echo as a whole. However, these storms, like
the LP supercell, all contain a mesocyclone. It should be noted that there are often attempts to type
supercell storms via visual appearance. However, because distributions of large raindrops and hail
are easily detected by radar, while visually these same hydrometeors often are not, these attempts

4-31
are very often in error and should not be completely relied upon. These three types of supercells
are depicted more clearly by radar.
Again, it is stressed that regardless of which supercell type is being discussed, the fact that a storm
is a supercell (i.e., it possesses a deep, persistent mesocyclone) is very important and fundamental.
As we shall see, this fact has enormous implications concerning the severe weather expected, the
typical lifecycle, and magnitude of storm severity.
4.7.3.1 Classic Supercell Storm. The supercell model was first conceived as a
persistent, quasi-steady state storm consisting of a single cell that may have a lifetime of several
hours while typically propagating continuously to the right of, and more slowly than, the mean
winds. Common characteristics of a mature supercell storm are echo elipticity, an echo diameter of
15 nm or more, a hook echo at times, and the presence of a mesocyclone. With much better
observations we now recognize that while these storms are persistent, they often are not quasi-
steady and they commonly have some multicellular characteristics.
4.7.3.1.1 Recognition of a Classic Supercell Storm. Development of a supercell

storm can be monitored through the use of the 4-panel/multi-panel or all tilts (used in AWIPS)
reflectivity display and velocity displays. Supercell development typically begins as a non-severe
cell in a multicell cluster storm with the reflectivities initially aligned vertically or even with a down-
wind slope. As the updraft strengthens the reflectivity echo core is displaced toward the updraft flank
and the adjacent reflectivity gradient strengthens. Sometimes, the reflectivity pattern will begin to
elongate, with the strongest gradient on the right and rear storm flank. The development and/or
existence of a weak echo region or a bounded weak echo region can be determined from the multi-
panel display of different elevation angles of Reflectivity products (Figures 4-17 and 4-18). With
sufficient radar resolution and with a storm sufficiently close to the radar, a supercell storm will
typically exhibit a bounded weak echo region that will last for an extended period of time, normally
until the collapse phase.
The storm top will shift from over the reflectivity echo core region to a position either above the
strong low-level reflectivity gradient or the WER/Bounded Weak Echo Region (BWER) as the
storm develops those structural features during transition from non-severe to severe (Figures 4-15,
4-16, and 4-17). These characteristics can be seen when comparing the lower elevation slices with
the higher elevations on the 4-panel reflectivity display. A four panel/multi-panel display of
different elevation angles of Storm Relative Mean Radial Velocity products (SRM or SRR) can
provide information on the existence of a mesocyclone. It will also allow the detection and
monitoring of the height of the strongest mesocyclonic shear and rotational velocities. Often, that
strong shear will first appear aloft and, with time, lower toward the surface. When, and if, that
most intense shear is seen to lower near the earth’s surface, tornadogenesis may be imminent.
Moreover, if the storm is within about 102 km (55 nm), the development of possible TVSs can be
monitored by the same multi-panel SRM or SWA velocity products. At other times high shears
within the mesocyclone will develop initially in low-levels or within the boundary layer, Trapp et
al. (1999).

4-32
When the environment is characterized by a linear or unidirectional hodograph with sufficient
shear, as the storm evolves further, splitting may occur with intense “mirror image” supercells on
the left and right flank. The low-level echo may begin to lengthen in a direction normal to the
storm movement as the splitting process begins. In addition, the echo aloft may split into separate
convective cores earlier than at the lower levels. Along with the 4-panel/multi-panel product,
continuous updates of Reflectivity products in a time lapse will aid in recognizing these trends.
The left flank storm typically has the WER/BWER on the forward left storm flank. Note the left-
moving supercell in Figure 4-16. This storm was the left-moving portion of a splitting storm and
the front flank WER and echo overhang are obvious in the 4-panel product. Left-moving
supercells typically move appreciably to the left of the winds in the cloud bearing layer and move
more rapidly than those winds. A deep, persistent anticyclonic circulation (“mesoanticyclone”)
signature is common with the “left-mover.”
Severe weather with the left-mover is typically large hail (sometimes > 5 cm (2 inches) diameter)
and damaging surface winds. Tornadoes are uncommon.
The right-moving counterpart that results from this splitting storm is typically a Classic supercell
that moves to the right of the environmental winds and often produces large hail, damaging winds,
and sometimes tornadoes. When the environmental hodograph is less unidirectional and veers
increasingly with height, the left-splitting storm is suppressed and the right moving storm is
favored.
Large values of VIL may be present with a supercell storm. Rapid increases in VIL values may
signify the initial stages of severe weather activity; rapid decreases in VIL values may indicate
the collapse phase of the storm. However, caution is recommended because gaps in the VCP
may artificially create abrupt and large decreases in VIL simply as an artifact of sampling. In
addition to the use of VIL, the M and TVS products can also be used along with vertical cross
sections (again, when properly placed). Mean Radial Velocity products may be used to
determine the flow patterns in and around severe convective storms. However, the SRR, SRM,
and SWV may all be used more effectively to monitor the kinematic storm structure and the
existence of mesoscale/storm-scale circulations. Storm Relative Motion products, with storm
motion removed, make mesocyclone signature identification easier and should be used to
ascertain a mesocyclone’s presence and rotational velocity.

4-33
Figure 4-16
Left-Moving Severe Thunderstorm Reflectivity
Norman, OK WSR-88D (test bed radar) 4-panel display of Reflectivity products at 22:53
UTC on 8 March 1992 (PUP display). This display shows a left-moving supercell
thunderstorm that developed from a storm split. The front flank Weak Echo Region and
echo overhang is revealed by noting the position of the geographically fixed cursor location
and the storm core as seen in this four elevation angle sampling of the storm. Storm
motion was 200° at 24.7 ms-1 (48 kts), which is over 10° to the left and over 9.3 ms-1 (18
kts) faster than the mean environmental wind.

4-34
Figure 4-17
Classic Supercell Thunderstorm Schematic
Schematic of a classic supercell storm with vertical cross sections along and at right
angles to storm motion. This cross section schematic shows a BWER as well as the WER,
echo overhang, and low-level hook echo. Note that while there is essentially no tilt in this
schematic, not uncommonly there is some storm tilt. From Chisholm and Rinick (1972).

4-35
Figure 4-18
Supercell Thunderstorm Reflectivity
Aberdeen, SD WSR-88D 4-panel display of Reflectivity products at 00:29 UTC on 24 June

2004 (AWIPS display). This display of a supercell storm has reflectivity slices at 0.5° in the
upper left, 2.4° in the upper right, 6.2° in the lower right, and 10° in the lower left. Note the
boundary across the lower portion of the upper left panel and the storm gust front trailing
from the storm hook echo intersecting. The low-level hook echo partially surrounds, with
height, a BWER.

4-36
4.7.3.1.2 Considerations. With a single Doppler radar providing velocity data,
only a portion of the flow in and around the storm will be seen. Moreover, velocity data can be
sampled only where and when sufficient tracers are available. Note, as well, that while the M and
MDA algorithms and associated products can be used with some success for mesocyclone and
supercell identification, these algorithms will not identify the mesoanticyclones of left-moving
supercells.
The storm reflectivity and velocity structure must be monitored continuously. This is done most
effectively through the use of the multi-panel reflectivity and velocity products during all
successive volume scans (Figures 4-16 and 4-19). Typically, supercells will have one major
storm top (often among other lesser tops) which will persist for a long period of time. Normally,
the SCIT algorithm will adequately identify and track the storm and provide storm structure
information. However, the forecast movement by the Storm Tracking algorithm will lag motion
changes. Very pronounced veering to the right will sometimes occur during supercell
development. Sometimes pronounced left deviation will occur during weakening.
4.7.3.2 High Precipitation Supercell. High Precipitation (HP) supercells

produce a variety of severe weather, including very large hail and occasional tornadoes, but they
also sometimes produce flash flooding (Moller et al. 1990). An HP supercell, like many
Classics, may exhibit some characteristics of both a multicell cluster and a supercell storm. As
mentioned earlier, the HP supercell typically is characterized by a front-flank updraft and
mesocyclone, as well as heavy precipitation to the rear and partially surrounding the storm’s
primary updraft (Figures 4-19 and 4-20). However, at times a Classic supercell can transition
into an HP as the mesocyclone (and a tornado, if present) becomes embedded in heavy rain.
Thus, for a short period, HP structure may result when a Classic supercell undergoes the collapse
phase. In addition, HPs are often characterized by a highly reflective and broad hook echo when
the hook is present. Visual observations indicate that heavy precipitation exists in areas where
the classic supercell is typically rain-free or includes only limited precipitation. These
characteristics sometimes make HP supercells more difficult to recognize.
4.7.3.2.1 Recognition of a High Precipitation Supercell Storm. The radar

reflectivity pattern of an HP supercell will often indicate a broad, highly reflective (over 50
dBZe) pendant or hook echo. Because HP supercells often have a front flank updraft, the WER
and BWER are generally located on the front flank. In these cases, the high precipitation
supercell has a front flank mesocyclone, in contrast to the right rear flank mesocyclone typically
observed with classic supercells. Occasionally, there may also be a central heavy precipitation
region with very strong or damaging outflow behind a bowing gust front. This region often has
sharp lateral edges to the damaging outflow and is typically to the right of the mesocyclone
center. It is not uncommon for these storms to produce very large hail and very damaging
surface winds. Occasionally, these storms are extremely severe or “high-end” events (Miller and
Johns 2000; Miller et al. 2002).

4-37
Figure 4-19
High Precipitation Supercell Thunderstorm
Oklahoma City, OK WSR-88D 4-panel display of products of a HP Supercell storm at 20:14

UTC on 11 May 1992 (legacy PUP display). The upper left display is 0.5° Base
Reflectivity, the upper right is a 0.5o SRM, the lower left is a 1.5° SRM, and the lower right
is a 2.4° SRM. The mesocyclone velocity couplet is near the white cross (109 km (59 nm)
from the radar) with inbound velocities being green and outbound velocities being red.

4-38
Figure 4-20
Reflectivity and Mean Radial Velocity Display of a High Precipitation Supercell
St Louis, MO WSR-88D 3-panel Reflectivity products and a single quarter-screen SRM

product of a HP Supercell storm (legacy PUP display) at 02:18 UTC on 8 September 1992.
Note the front flank notch (upper left) and the mesocyclone in the lower right SRM product.
The mesocyclone and TVS velocity signatures (right panel) are embedded within moderate
to heavy rain as indicated by the surrounding (red) reflectivity echo.

4-39
A low-level Reflectivity product will reveal the inflow notch on the side of the mesocyclone
characterized by inflow and updraft. Again, with the HP, this is typically on the storm’s front
flank. The weak echo region, storm overhang, and highest storm top are typically on the leading
edge of the storm. This is best monitored via the multi-panel Reflectivity and Velocity (V, SRR,
SRM, and SWV) displays.
The existence of a mesocyclone may be apparent in an SRM or SRR product, a Mean Radial
Velocity product, or, to a limited extent, the VCS product oriented perpendicular to a radial
through the storm. As always, placement of the vertical cross section is critical and the meanings
of the velocities are much more difficult to determine when not radially aligned. Heavy rainfall is
easily identified with the OHP product. In fact, especially with HP supercells, the precipitation
accumulation products OHP, THP, STP/DSP, and the User-Selectable Rainfall Accumulation
Precipitation (USP) are of critical importance. When applying the WSR-88D, the user must be
careful not to neglect certain potential threats (e.g., flooding) when dealing with others (e.g.,
tornadoes, wind, and hail). If an event is not anticipated, when it occurs it is often not recognized
in a timely manner (Quoetone and Huckabee, 1995).
Significant changes in a high precipitation supercell, such as rapid transition to a bow echo with a
rotating comma head, or movement of a mesocyclone to the northeast within the echo mass, can be
monitored through the use of Reflectivity, and Velocity product time lapses. Climatologically high
VIL values, large accumulations of rainfall, and high values of reflectivity are likely with High
Precipitation supercells. Time lapse sequences of VIL, precipitation accumulation, and
Reflectivity products should be extremely useful in monitoring storm evolution and in anticipating
a transition from an HP into a Bow Echo.
4.7.3.2.2 Considerations. Due to the amount of precipitation, front flank

mesocyclones, and broad embedded hook echo structures of high precipitation supercells, they are
sometimes difficult to identify on radar. At other times, these structures are prominent making the
HP more easily identified. Moreover, the SCIT algorithm may be unable to adequately track an HP
supercell when the storm is embedded within considerable surrounding strong convection such as
within a Mesoscale Convective System (MCS). At times, storm components may be interchanged
between adjacent echoes. This can create abrupt changes in tracking forecasts and can impact the
output of other algorithms such as the hail algorithms. This underscores the importance of doing
base data analysis with decreased reliance on algorithm output. The time lapse capability of base
data can be useful in monitoring a storm's evolution and changes in motion. But the multi-panel
displays of base data are critical in monitoring storm structural changes and storm intensity/severity
changes via the three-dimensional echo structure.
Microbursts are common with Classic and HP supercell storms, especially in the vicinity of the
hook echo, mesocyclone, RFD, and front flank notch. In fact, the RFD actually promotes them.
However, because the most intense and damaging outflow from microbursts commonly occur
within about 300 meters of the ground, they are very difficult to detect via radar. Moreover, they
are usually not axis-symmetric, thus they may be even more difficult to detect depending on the

4-40
viewing angle. Base velocity data with high resolutions are preferred, as is use of VCP 12, which
includes increased and overlapping low-level sampling over previous VCPs.
4.7.3.3 Low Precipitation Supercell Storm. Low precipitation supercell storms

are rather rare, and to the limits of our knowledge, virtually unique to dryline environments within
the Great Plains and High Plains east of the Rocky Mountains. While appearing benign in
reflectivity fields and even in velocity fields at times, low precipitation supercells frequently
produce large hail, but rarely wind damage, because of the lack of a significant downdraft. While
damaging tornadoes can be produced, they are rare with the LP supercell (Doswell and Burgess
1993) (Figure 4-22).
4.7.3.3.1 Recognition of Low Precipitation Supercell Storms. Radar reflectivity

of 50 dBZe or less is common with low precipitation supercells. Occasional higher reflectivity
caused by hail at mid-levels of the storm may be detected at higher elevation angles of Reflectivity
products. Storm Tracking Information products typically function well with LP storms because
they tend to be more isolated than HPs, for example. In addition, the Hail Detection Algorithm
(HDA) may provide data that will allow some assessment of the hail potential.
Due to the sparsity of rainfall, low precipitation supercells rarely contain a pendant or hook echo,
especially when the storm is beyond close radar ranges (Figure 4-22). The mid-level mesocyclone
can often be identified as with other supercell storms through use of multi-panel displays of V or
SRM, SRR, or a series of 4-panel Severe Weather Analysis (SWA) products. With these products
the user can inspect the displays for symmetry, depth, and rotational velocities as a function of
range from the radar. But with less precipitation sometimes detection of the mesocyclone and
other structural details are more difficult. Further, mesocyclone detection may be confined to
storm mid-levels. Moreover, since LP storms typically exhibit few, if any, distinctive reflectivity
features and, since detection of the classic echo features is so range- and resolution-dependent,
these storms are often difficult to identify, especially without adequate velocity field examination.
As always, Velocity products may be used to determine the flow patterns in and around LP
supercell storms. However, SRM products, with the correct storm motion removed, make
mesocyclone signature identification easier for the user and should be used to determine
mesocyclone shear and rotational velocity. But, because shear is invariant to storm motion,
algorithm detection does not depend on storm motion removal.
4.7.3.3.2 Considerations. Reflectivity structures of low precipitation supercells

may be much more difficult to resolve because of the scarcity of precipitation and their typically
small storm size. In addition, the lack of significant reflectivity may cause a low precipitation
supercell to be overlooked, or its strength underestimated for most radar distances.
Significant hail formation within a low precipitation supercell can adversely affect the Precipitation
Processing algorithms. These storms will frequently produce 0.2 inches or less of rain at locations
along their paths. But with hail, radar estimates of precipitation will often be somewhat greater.

4-41
4.7.3.3.3 Mini-Supercells. During the past decade, supercelluler convection has
been observed to take place on much smaller scales than the larger counterparts just discussed
(Figure 4-21). So called “mini-supercell” storms are typically much smaller in all dimensions.
These miniature supercells have been discussed by McCaul (1987); Burgess et al. (1995); Grant
and Prentice (1996); and others. While these storms appear to be small, they still can, and do,
produce significant and sometimes substantial severe weather, including tornadoes. Indications are
that in mid-latitudes, the environment for at least some of these storms appears to be marginal
Conditional Available Potential Energy (CAPE) but strong shear. In other cases, the CAPE and
steep lapse rates are confined to a limited, but lower-level depth, of the atmosphere. (This is also
the case with “low-topped” supercells which are very similar to the mini-supercell, except that
other dimensions are somewhat larger.) Moreover, it appears that most of the storms producing
tornadoes associated with land-falling tropical storms are of the mini-supercell variety (Figure 4-
47). In these cases much of the environmental shear is confined to the lowest few thousand feet
above ground.
However, because of the small dimensions, low-tops, and small mesocyclones, these storms can
only be resolved when relatively near the radar. The 4-panel Reflectivity and Velocity products are
still appropriate, but high resolution products are advised. It will also be appropriate to use VCP
12, with its concentration of sampling via overlapping beams in lower levels.
4.7.4 Severe Thunderstorm Identification. Of great importance are the characteristics of

a thunderstorm radar echo that indicate the storm is severe. Of equal importance are those
characteristics that indicate a transition from a non-severe to a severe thunderstorm. Analysis of
the three-dimensional structure of a thunderstorm cell provides fundamental insight into the
severity of a storm. For example, the presence of a WER or BWER and strong storm top
divergence are important indicators of an intense updraft and storm severity. When mid-level
reflectivity is strong, the storm WER is large, and the storm echo top is over the WER, the
likelihood of at least a severe hailstorm is virtually certain (Lemon 1980; Lemon and Doswell
1979; Burgess and Lemon 1990; Moller 2001). The presence of a BWER enhances the probability
of a thunderstorm producing a tornado. It is these and other features we will address here.
4.7.4.1 Storm Structure. We have been, and will be, speaking of the “updraft
flank” or “storm inflow flank.” The updraft storm flank in the northern hemisphere is often to the
right or right rear relative to storm motion. However, relative to squall lines, the updraft storm
flank is normally along the advancing or leading edge (forward flank) of the line. In reality, the
flank of the storm where the updraft is located is unimportant; it can be any flank. But the echo
characteristics we address below are independent of storm flank and, to a great extent, storm type,
as well. It is not the storm flank that is important, rather the characteristics of the updraft
positioned there.

4-42
Figure 4-21
Mini-Supercell Thunderstorm Reflectivity
Fort Drum, NY WSR-88D 4-panel Reflectivity product display at 20:53 UTC on 3

September 1993 (PUP display). A mini-supercell storm (near the white cross) is 16 nm
from the radar. These Reflectivity products are slices with the following elevation angles:
1.5° upper left, 4.3° upper right, 5.3° lower left, and 7.5° lower right. The low-level pendent
echo partially surrounds a BWER with increasing height in this very small and shallow
storm.

4-43
Figure 4-22
Low Precipitation Supercell Thunderstorm Reflectivity
Fort Hood, TX WSR-88D Reflectivity product at 00:05 UTC on 22 March 2005 (NCDC
graphics). This LP Supercell storm located in Falls and Limestone counties of central
Texas produced a weak tornado and hail.

4-44
Storm severity is most often a function of updraft strength. The stronger and more intense the
updraft, the more severe the storm is (Burgess and Lemon 1990). Thus, radar is used to diagnose
updraft trends or strength. In a “pulse” severe storm or multicell storm environment, an indication
that there is a transition to severe activity is the appearance of a cell with its first echo at a higher
level than the earlier cells owing to an invigorated updraft. Viewing higher elevation angles of a
Reflectivity product, especially by using the multi-panel display, can aid in determining this.
However, this is often very difficult to recognize in real-time.
There are other more obvious and more significant changes that take place in a radar echo within
an environment of deep layer shear and moderate to strong instability that signal the onset of a
strengthening updraft and a severe storm (Lemon and Doswell 1979; Lemon 1980; Browning 1977;
Burgess and Lemon 1990). We have previously covered a portion of this material in Section 4.7.3
but it is reviewed and summarized here because of its importance.
What are these changes or echo characteristics of a strengthening updraft in a sheared

environment? First, as illustrated in Figures 4-15 and 4-18, in the low-level echo, a tightening
reflectivity gradient will occur on the storm inflow and updraft flank. At times, this storm flank
will also develop a concavity open to low-level inflow in association with the strong reflectivity
gradient. A second related change is that the dominant low-level reflectivity echo core is typically
displaced toward this storm flank.
With increasing storm summit divergence and rapid expansion of echo aloft, a mid-level echo
overhang develops and extends over and beyond the strong low-level reflectivity gradient,
producing a WER. The overhang and WER can be identified by comparing several elevation
angles of Reflectivity in the multi-panel display, or with an RCS that is well-placed. However, the
first choice is the 4-panel or multi-panel display as explained previously. At the same time the
storm VIL will be increasing, reflectivity at all levels will be strengthening, and hail algorithm
probabilities (Probability of Hail (POH), Probability of Severe Hail (POSH), and Maximum
Expected Hail Size (MEHS)) will be increasing. Depending on the VCP and other scanning
characteristics, the storm echo top may be characterized as increasing.
Within the developing echo overhang associated with the WER is development of a strong mid-
and upper-level reflectivity echo core, again located above the tight low-level reflectivity gradient.
As updrafts have become stronger, the echo top shifts over the tight low-level reflectivity gradient
or above the WER on the updraft storm flank. Note that it is not the height of the echo top that is
important, it is the location of the top relative to the low-level echo. A multi-panel velocity product
during this early phase of supercell development may begin to show a mid-level (often cyclonic)
shear signature, but it is not of sufficient strength as of yet to indicate a mesocyclone.
Storms growing within an environment of deep layer shear and moderate to strong instability that
develop this structure are severe storms. The dominant severe weather phenomena these storms
produce are large hail (> 1.9 cm (¾ inch hail)) and significantly less often damaging surface winds
(> 25.7 ms-1 (50 kts)).

4-45
4.7.4.1.2 Recognition of Severe Storm Structural Features. As emphasized
within this document the most useful technique in analyzing the structure of a storm is the
generation of multi-panel Reflectivity and Velocity products (V, SRR, and SRM) with
progressively higher elevation angles and magnification on the storm in question (e.g., Figures 4-
18, 4-19, 4-29). This can often reveal the WER and echo overhang resulting from intense storm
summit divergence. At times, if generated correctly, an RCS product will also show a strong
midlevel reflectivity field and its overhang position creating the WER below, again indicating the
presence of intense divergence from a strong updraft. Evidence of intense upper level storm
summit divergence (an indication of updraft strength) can often be found through the use of a
higher elevation scan of the SRR, SRM, or V products and, occasionally, through careful
placement of a VCS product. (The overhang must extend horizontally for few miles to be
significant.) This region beneath the overhang, the WER, should be easily identified.
The development of a conically shaped BWER extending vertically upward into the echo overhang
base typically indicates an intense updraft (Figures 4-18, 4-21, and 4-23). This feature may
develop in response to intensifying updraft or due to improved radar resolution, or both,
because of decreasing range to the storm. In a mid-level Reflectivity product, a BWER will
appear as a hole, a circular or elliptical region of low reflectivity values enclosed by higher values.
At times, the BWER will be elongated and curved with the long axis parallel to the low-level gust
front where the intense lifting may originate. While the BWER should be evident in an RCS
product if properly placed, its presence and development is again best detected in a multi-panel
Reflectivity product with magnification (Figure 4-18). Evidence of associated mesocyclone
development may be found in a companion multi-panel velocity (V, SRR, and SRM) product,
again with magnification (Figure 4-19). Most often, the mesocyclone makes its initial appearance
in mid-levels near or within the area of the BWER.
In addition, use of the multi-panel displays can also be used effectively to determine if the storm
top location is very near or over the WER, BWER, or the strong low-level reflectivity gradient.
Within the supercell life cycle, the collapse phase is critical. It often is most easily recognized in
reflectivity structural features and is critical for severe weather occurrence at the ground, (Brown et
al. 1973; Lemon and Doswell 1979: Burgess et al. 1982; Burgess and Lemon 1990). The wrap-up
and disappearance of the hook echo, collapse or diminishment of the classical reflectivity features
of BWER, WER, storm top descent, and a rapid decrease in the VIL values will often signify the
decrease in surface hail size, but will also signal the onset of surface wind damage (and
microbursts) and tornadoes. It is during this same time period that tornadogenesis is often typified
in the velocity data by the descent and surface arrival or low-level intensification of the
mesocyclone shear, rotational velocity, along with circulation diameter decrease. The development
of the low-level TVS (Figure 4-29, Figure 4-31, and Section 4.7.4.4), when within radar range (<
102 km (55 nm)), is often detected at this time. Because of the importance of this storm phase the
user must carefully monitor reflectivity and velocity displays. However, detection of these features
near the earth’s surface is heavily dependent on the radar range to the storm and related radar
horizon.

4-46
Figure 4-23
Mature Severe Thunderstorm Schematic
Same as Figures 4-12 and 4-15 except for a mature Classic supercell storm. Note the
position of the BWER and storm top as is typical of these storms in the mature
development phase. After Lemon (1980).

4-47
For a more detailed look at severe storm and supercell morphology and structure the reader is
directed to Lemon and Doswell (1979); Lemon (1980); Burgess and Lemon (1990); Doswell and
Burgess (1993); and several chapters within Doswell (2001).
4.7.4.1.3 Considerations. Alphanumeric data from the Storm Structure algorithm

and the associated “trends” displays of some WSR-88D display systems can provide information
on the structure and evolution of a storm. However, reflectivity and velocity product 4-panel
displays may be more useful in determining the three-dimensional structure of a severe storm.
Understanding the characteristics and model of a storm class is invaluable in determining its
potential surface severity and lifecycle stage. Monitoring the history of storm tops (development,
location, and collapse), the intensity, and altitude of the maximum reflectivity, and other such
parameters, can provide insight into where a thunderstorm is in its life cycle and the surface severe
weather production. These insights may be gained through monitoring of the storm velocity and
reflectivity structure. As before, multi-panel base product display is preferred for mid- and upper-
levels, as well as the lowest elevation angles in order to monitor storm structure and mesocyclone
development and strength. Vertically Integrated Liquid, ET, and high-level LRM may also be used
to monitor storm significance. Time-lapse loops of 6 to 10 frames are often most useful.
Additionally, the user must be careful to discern a real weakening and diminishment of the
supercell storm from the collapse phase. This can be done by coupling the velocity information
with the reflectivity field information. Unless the WSR-88D user is careful, noting the diminishing
reflectivity structure can lead the warning meteorologist to conclude that the storm is no longer
dangerous when, in fact, the storm is entering its most dangerous and deadly phase.
4.7.4.2 Echo Shape, Severe Weather, and Damaging Winds. During the
early years of radar development and its application to severe weather detection, echo shape
was the radar practitioner’s “first line of defense.” The shape of a radar echo was viewed as
an indicator to thunderstorm severity. We have since recognized that many of these shapes
were highly subjective, did not rely on our physical understanding of storms, and often had a
rather high false alarm rate. However, to some extent, we still do rely on echo shape as a
clue to severe weather potential.
For example, hook echoes have long been associated with tornadoes and are often indicative
of supercell storms, although tornadoes do not always accompany them (Figures 4-18 and 4-
24). Also, a tornado may be well developed without a radar hook being present. Hook echo
detection and recognition is also a function of radar range and resolution. Beyond about 60
km (100 miles) true hook echo detection is rare. Moreover, during tornadogenesis and
during an increase in tornado strength the hook echo will often disappear as the other echo
structural features weaken and vanish as the collapse phase of the storm proceeds. Thus, a
hook echo alone has been found to be of questionable value. However, actual supercell hook
(or pendant) echoes must be, or have been associated with, a WER or BWER. Thus, the
overall storm structure is important in verifying such reflectivity features. Moreover, the
detection of an associated mesocyclone is even better confirmation.

4-48
Figure 4-24
Tornadic Supercell Thunderstorm Reflectivity
Oklahoma City, OK WSR-88D Reflectivity product at 03:49 UTC on 10 May 2003 (AWIPS
display). A supercell with a hook echo is north of the radar. A strong tornado was
occurring within the hooked or curl-shaped echo appendage on the southwest portion of
the storm located north-northwest of the radar.

4-49
Echo shape has especially been useful as severity indicators of thunderstorm lines. For example, a
thunderstorm Line Echo Wave Pattern (LEWP) may develop, then a portion will take on a bow
echo shape. These bow echoes will occasionally develop a “rotating comma head.” Like the
LEWP, the bow echo is often accompanied by severe weather at the surface such as damaging,
low-level diverging winds with macrobursts or microbursts. The rotating comma head is often
indicative of a mesocyclone and, at times, may be accompanied by tornado development.
Przybylinski and Gery (1983) reported that “distinctive” bow echoes are convex in shape, with a
strong, low-level reflectivity gradient along the leading edge, the maximum echo top displaced
over or ahead of the gradient, and an evident weak echo channel. Note that most of these echo
features are the same radar echo features emphasized previously that mark an intense updraft and
severe weather-producing storm. Thus, Przybylinski and Gery (1983) offered further confirmation
of the importance of these echo features and shapes.
4.7.4.2.1 Recognition of Echo Shape and Attendant Features. A low-level

Reflectivity product will show the presence of bow- and comma-shaped echoes (Figures 4-25 and
4-26) and attendant features, but the three-dimensional storm structure should also be viewed along
with the low-level echo. The 4-panel/multi-panel reflectivity and velocity displays can be used to
verify the impression of low-level echo shape. Moreover, the output of the MDA via the
Mesocyclone product or overlay, and the TDA via the output of the TVS or overlay, can also be
used for further verification. However, the output (positive or negative) of these and other system
meteorological algorithms must always be verified by examination of the velocity products. The
maximum echo top location can be identified from the ET and EET products. However, high-
levels will provide better resolution of where the echo top is located and its height. A comparison
can be made as to top location in relation to the low-level reflectivity gradient by displaying the
products on a multi-panel display and linking the cursors to permit visual comparison of low-level,
mid-level, and high-level echo features. Reflectivity Cross Section products can also help in
identifying the low-level gradient and the maximum storm top position, as well as the general
storm structure, but must be positioned properly.
With squall lines and bow echoes, the “Rear Inflow Jet” (RIJ) and low-level weak echo channel
shapes (Rear Inflow Channel, RIN) to the rear of the echo line have been emphasized as being
associated with damaging winds (Przybylinski 1995). Within the limits of the viewing angle,
strong velocity maxima near these RINs should be identifiable and used with low-level Velocity
products for verification. The weak echo channel is often located near the center of the bow and
associated with the RIJ and strongest surface downburst winds (Figures 4-25 and 4-26). (See Part
B, Chapter 7, of this Handbook for discussion of squall lines, MCSs, and bow echoes.)

4-50
Figure 4-25
Bow Echo Schematic
Conceptual model of bow echo characteristics and typical evolution. Adapted

from COMET (1999).

4-51
Figure 4-26
Squall Line Schematic
Development of cyclonic (north) and anticyclonic (south) “bookend” vortices in a squall line
simulation. Note the Rear Inflow Jet. From the COMET MCS module (COMET, 1999).

4-52
A time lapse of a, specifying a continuous update, will aid in monitoring squall line feature
evolution such as bow- and comma-shaped echoes. Sometimes bow echo development may be
noticed first with the appearance and acceleration of a forward bulge in the echo or the line and
sometimes with the development of a LEWP, a portion of which can later assume a bow shape.
The bow echo may further evolve into a comma-shaped echo with a rotating head. Note that
sometimes the most damaging winds associated with a Bow Echo or LEWP occur during
development of these features. Once the bow shape is well developed the strongest wind damage
may be past. The comma head is a mesocyclonic storm, i.e., a supercell, even though rotation may
be confined to mid-levels. Rotation at the head of the comma echo may be identified in the SRM
or SRR products, the Mesocyclone product, and perhaps in a time lapse of Reflectivity products.
Important three-dimensional features of damaging Bow Echoes are: elevated RIJs, line-end
vortices, rear inflow notches, and the indicators of an intense updraft as given earlier: strong
reflectivity gradient, WERs, displaced echo top, and supercells. The line end vortices can be very
important because they can amplify the RIJ and can be associated with tornadoes at their leading
edge.
Increasing velocities behind a line echo (or large echo) may signal acceleration of a portion of the
line and transition into a bow shape. Very damaging winds and some large hail may occur during
this phase. Bow-shaped echoes, such as within a LEWP, may move at a faster speed than the line
itself owing to the rear inflow jet. If other storms do not intersect its path, or are not in close
proximity, the bow echo's movement should be effectively tracked by the STI product.
The appearance of a low-level hook echo and/or the “inflow notch” along the leading edge of a
squall line, comma head, or bow echo can best be identified via the three-dimensional echo
structure detected in reflectivity and velocity products through the multi-panel base product
display. Again, a time lapse may aid in monitoring the evolution of this feature. However, rapid
changes in mesocyclone and storm strength are not uncommon and the user must be
knowledgeable of the near storm environment in order to evaluate the need for short fuse tornado
or damaging wind warning. The rotation of the “bookend vortices” should be noticeable in
velocity products (V, SRR, SRM) in mid- and/or low-levels. One of the indications of enhanced
tornado threat with linear echoes is the intersection of the line with a surface-based boundary, often
located at right angles to line orientation (Przybylinski et al. 2000). This threat is further enhanced
with associated hook echo development. Another sign of enhanced tornado threat is the
development of strong echoes just ahead of the line and the subsequent intersection of that echo
and the line itself.
4.7.4.2.2 Considerations. Windstorm identification relies heavily on echo shape,

but other factors need to also be considered. For example recognition of storm-damaging wind
potential relies on the LEWP, on bowing line segments, the surging supercell RFD, detection of the
DCZ or the MARC, mesocyclones, WER, BWER, low-level tight reflectivity gradient along the
leading edge of a squall line, and the rear inflow notch.

4-53
Severe weather can be produced during bow echo or comma echo development. Since it is often
hard to identify individual echoes assuming a bow shape in real-time, the use of time lapse with
continuous update of a Reflectivity product can be extremely useful in increasing warning lead-
time. Storm structural features such as the WER, BWER, hook echoes, inflow notches, and echo
shapes, such as the Bow Echo, often cannot be distinguished in a CR product since their shapes are
low- and mid-level features and are obscured in CR products by reflectivity echo canopies
(overhangs) and cores above. However, the CR Combined Attribute Table provides an excellent
summary of algorithm output.
Range-overlaid echo may obscure velocity data needed to confirm a mesocyclone or TVS, should
they exist, but not the reflectivity hook echo itself if it exists. Sometimes, however, the higher
elevation velocity data may be free of overlaid echo. Moreover, velocity signatures should be
interpreted in the context of the storm reflectivity structure.
As a review, damaging wind radar patterns/trends include LEWPs, bowing segments, surging
RFDs, deep convergence on backside of mesocyclones, the DCZ and the MARC, mesocyclones
(non-descending, descending, developing, strengthening), WER, BWER, tight reflectivity gradients
along the leading edge, the RIN, and gust front speeds well matched to squall line speed. Descent
of high reflectivity cores and associated mid-level convergence are also indications of downbursts
and/or microbursts.
Types of radar products used for analyzing potential damaging wind storms include high resolution
R/V, R/SRM multi- or 4-panels, all tilts, cross sections, SW, and time-lapse animations. However,
keep in mind that the SRM does not display ground-relative velocities and should not be used for
surface wind warnings.
4.7.4.3 Mesocyclone Signature Detection. A mesocyclone is a small scale

rotation closely associated with a major convective storm updraft (in which vertical velocity and
vertical vorticity are correlated) and often involves the RFD, as well. True mesocyclones meet or
exceed established thresholds for: persistence, (minimum of two volume scans), vertical extent
(shear extends at least 3 km (10,000 ft) in the vertical, but dependent of storm depth), and shear.
The distance between maximum inbound and maximum outbound velocity peaks of the
circulations are normally ≤ 9.3 km (5 nm). The rotational velocity is equal to the absolute values of
the maximum inbound velocity plus the maximum outbound velocity divided by 2 (using mid-
range velocity product values). While the radar system is equipped with algorithms (the legacy
Mesocyclone algorithm and the recently developed MDA) for the automated identification of these
features, here we confine our discussion, for the most part, to manual identification and
meteorological meaning of these features. The algorithm designs are discussed in Part C, Chapter
3, of this Handbook.
An overwhelming majority (probably more than 95%) of mesocyclones (Figures 4-27 and 4-29)
have been found through long-term Doppler observations to be associated with some form of
severe weather (hail > 1.9 cm (¾ inch), winds > 25.7 ms-1 (50 kts), and/or tornadoes). Perhaps

4-54
26% are precursors to, or associated with tornadoes (Trapp et al 2005). They generally have the
size, depth, and persistence characteristics just mentioned. Mesocyclones have been found to be
associated with virtually all strong and violent tornadoes. Supercells (storms containing deep
persistent mesocyclones) are responsible for a disproportionate amount of damage. Stronger
mesocyclone signatures more often produce tornadoes than do weaker signatures, especially if a
TVS and BWER are also present. However, as of this writing, we have yet to be able to effectively
identify or discriminate via radar between mesocyclones that will produce tornadoes from those
that will not.
4.7.4.3.1 Recognition of a Mesocyclone Signature. A mesocyclone typically will

develop within mid-levels of a supercell storm and build upward and downward as the storm
matures. Mesocyclone rotational velocity, shear heights, and shear strengths can be monitored via
the velocity field revealed in multi-panel V, SWV, SRM, and SRR products. The M, MDA, and
TDA algorithm products, though they have relatively high false alarm rates at times, can also act as
a “safety net.” But the output of these algorithmic products must always be verified with velocity
products. Time lapse capability with continuous update may also be useful, at times, in monitoring
mesocyclone development. As a first approximation the velocity field of a mesocyclone will
appear as a solidly rotating core (Lemon et al. 1977) (Figure 4-28). That is, the mesocyclone
appears as two velocity peaks of opposite signs, separated azimuthally (Figures 4-27, 4-28, and 4-
29). The surrounding potential vortex (outside of the core circulation) is more difficult to discern
with radar.
Interrogation of different elevation scans of SRM or SRR products (e.g., Figure 4-19, at 0.5°, 1.5
km (5,000 ft); 1.5°, 3.6 km (12,000ft); and 2.4°, 5.5 km (18,000 feet)) can aid in determining the
height continuity of a mesocyclone. As seen in Figure 4-29, the multi-panel mode can be very
useful for this purpose.
If correctly oriented through the storm, storm rotation (cyclonic or anticyclonic) of a well-
developed mesocyclone may be evident in the RCS product. Remember, however, that correct
placement is critical. When the velocity cross sections are not oriented along a radar radial, then
interpretation of these velocities displayed in the product become problematic. Therefore, for
operational purposes, the multiple panel R, V, SRM, SRR, and SWA products become far more
useful to the user during storm interrogation.
Removing the storm motion via SRR or SRM products may aid in the human recognition of a
mesocyclone when shifted velocities caused by this storm motion make recognition difficult.
Algorithm detection is, however, unaffected by this storm motion.

4-55
Mesocyclones
Figure 4-27
Mesocyclones
Oklahoma City, OK WSR-88D SRM product at 03:01 UTC on 4 May 1999 (AWIPS
display). Three mesocyclone velocity couplets (encircled) are north-northwest of the radar
and are identified by the bright green adjacent to the bright red color-coded velocities.

4-56
Figure 4-28
Combined Rankine Vortex Schematic
A schematic of the first approximation of a mesocyclone showing the velocity profile as a

Combined Rankine Vortex. From Lemon et al. (1977).

4-57
Figure 4-29
Mesocyclone and TVS Vertical Extent
Birmingham, AL, 6-panel WDSS Storm-Relative Mean Radial Velocity (SRV on WDSS
display) depicting WSR-88D SRR product for 6 progressively higher elevation slices. A
mesocyclone and TVS associated with the supercell and tornado that struck Birmingham,
AL on 8 April 1998 are depicted. The radar is in the lower right. Outbound velocities are in
red and inbound in green. The increase in color brightness indicates progressively higher
velocities. In the upper left the height is 1.3 km (4,264 ft) above radar level (ARL), upper
center, 2.8 km (9,184 ft), upper right, 4.3 km (14,104 ft), lower left, 6.2 km (20,336 ft), lower
center, 8.1 km (26,568 ft), lower right, 11.3 km (37,064 ft) ARL. A tornado was in progress
as these data were collected.

4-58
In a tornadic supercell, Doppler data have indicated many times that a separation of the
mesocyclone core from the BWER occurs prior to or during the collapse of the BWER and often
just prior to tornadogenesis (Lemon and Doswell 1979). With the somewhat increased likelihood
of tornado development at this stage, the separation of the mesocyclone core from the BWER may
be monitored via the multiple panel displays.
However, mesocyclone presence alone will not give adequate information to anticipate
tornadogenesis. Nor can we sample the buoyancy of the RFD (a positively buoyant RFD is
favorable to tornadogenesis). When the Lifting Condensation Level (LCL) is low (less than ~ 1000
m) and the 0 to 1 km shear is high (greater than ~ 25.7 ms-1 (50 kts)) these conditions then aid the
storm in tornado development and suggest a buoyant RFD may be present (Carven et al. 2002; and
Brooks and Craven 2002). When evidence is present indicating a source of low-level vertical
vorticity and a coexistence of strong low-level updraft, beneath or with this mesocyclone, then
tornadogenesis becomes even more likely (for more detail and discussion see
http://www.wdtb.noaa.gov/courses/dloc/ic57/ic57.pdf). As before, the use of high resolution
multiple panel base products provide the tools to monitor indications of tornadogenesis. While not
mentioned, magnification of these base product images is essential.
4.7.4.3.2 Considerations. The legacy Mesocyclone algorithm and the new MDA
provide the position of the feature on the lowest elevation angle in which the feature was detected.
In this way the meteorologist/user can better warn those beneath or within the mesocyclone
circulation. However, mesocyclones (and other storm features with a vertical extent) are often
tilted and are, therefore, displaced at higher elevations from where they are located in lower levels
and at the surface. Thus, both radar horizon and vertical slope must be considered before
concluding where a mesocyclone or a TVS may actually be located at the earth’s surface.
Mesocyclones are present with supercell storms and often produce severe weather and tornadoes,
but severe weather and tornadoes can also be associated with non-mesocyclonic storms. For
example, ordinary multicell cluster storms can also produce large hail, downbursts, or weak
tornadoes produced as a result of pre-existing vertical vorticity along convergent boundaries.
When present, and through the use of the MDA, mesocyclone tracking information is now supplied
to the user.
However, mesocyclones can, and often do, propagate discreetly with one mesocyclone core
dissipating while a second develops in a new but nearby location. In fact, about 40% of supercell
storms produce multiple mesocyclone cores during their extended lifetimes and can be “tornado
family producers” (Figure 4-30). Because of this fact, and because supercells often produce
tornadoes during the storm’s collapse phase, warnings for these storms should persist for a
significant period of time, even when they appear to have lost their supercell organization and
storm severity (Burgess and Lemon 1990).
In order to best resolve and characterize velocity-related storm features such as the mesocyclones
and TVSs, use of the Digital Velocity Array (DV) with 0.25 km (0.13 nm) range bins and 256 data
levels should be used during severe convective storm interrogation.

4-59
Figure 4-30
Multiple-Core Mesocyclonic Storm Schematic
Conceptual model of mesocyclone core evolution. Thick lines are low-level wind
discontinuities, and tornado tracks (exaggerated for size indication) are shaded. Insert
shows tornado family tracks and the small square is the region expanded in the figure.
From Burgess et al. (1982).

4-60
Due to beam broadening and radar horizon, Doppler radar detection of mesocyclones is generally
limited to about 230 km (124 nm). Mesocyclones range in size from 1.8 to 9 km (1 to 5 nm) in
diameter. “Misocyclones” formed in weak vertical wind shear may not be detected much beyond ~
46 km (25 nm) due to their small size ~ 500 m (1600 ft) to ~ 1500 m (4900 ft), (Lee and
Wilhelmson 1997).
Mesocyclones most often imply generation of RFDs (the LP storm may be an exception) and
strong updrafts. In order to increase confidence in their existence, they must have time and
height continuity within the data. Mesocyclone diameter is not related to severity of the
thunderstorm but, a shrinking mesocyclone diameter or increasing rotational velocity in low
levels should be interpreted as signs for significantly increased tornado likelihood. The user
must also monitor the data for low-level convergence with a mesocyclone above (or nearby)
implying vertical stretching of vorticity and tornadogenesis. In these very dangerous situations
the mesocyclone need not lower to the surface prior to tornado occurrence. In fact, when an
elevated mesocyclone is situated over strong low-level convergence the user should anticipate
that tornadogenesis is imminent (Burgess and Magsig 1998).
4.7.4.4 Tornado Vortex Signature Detection. The TVS is defined as a locally

intense circulation indicated by strong shear on the order of 0.01 s-1 between two velocity gates
which are azimuthally adjacent and constant in range (gate-to-gate) (Brown et al. 1978). It is an
intense gate-to-gate azmuthal shear associated with tornadic-scale rotation (Figures 4-29 and 4-31).
A TVS is identified manually if the gate-to-gate velocity difference is: > 46 ms-1 (90 kts) and the
range to the feature is < 55.5 km (30 nm). But when its range is between 55.5 km and 102 km (30
nm and 55 nm) the gate-to-gate velocity difference must be > 36 ms-1 (70 kts) to qualify as a TVS.
The gate-to gate shear is equal to the velocity difference between the maximum inbound velocity
value and that of the maximum outbound velocity (absolute values), normally in adjacent azimuths
and at the same range. These values are only guidelines; the user will have to adjust according to
the situation, storm size, near-storm environment, and geographic or topographical location. While
the system uses an algorithm for the automatic detection of the TVS, we confine our discussion for
the most part to the feature and its meteorological meaning.
Precisely what causes the TVS signature within Doppler radar data is unknown. We do know
that the radar beam is normally substantially larger than even the largest tornadoes even at
close ranges. What is also known is that the signature is highly correlated to vortex existence
of tornadic intensity, either incipient or actual tornadoes (Brown et al. 1978). Speculation
concerning the signature cause has included an “intermediate” circulation immediately
surrounding the tornado called the “tornado cyclone” (Burgess et al. 2002).
The hook echo, which has, in the past, been considered indicative of tornadic circulations, is,
in fact, more indicative of a mesocyclone. While tornadoes are not directly observed with even
Doppler weather radar, a TVS within velocity products is often a strong indication of presence
of a tornado. This is especially true if located in the hook echo and mesocyclone.

4-61
Figure 4-31
Tornado Vortex Signature Vertical Extent
Oklahoma City, OK WSR-88D 4-panel display with products from 3 May 1999 (WDSS
display). The upper left display includes a Tornado Vortex Signature Product (TVS)
overlaid on the reflectivity display. The remaining 3 panels contain SRR products. The
TVS signature extends from 1.67 km (5500 ft) above ground level (AGL) to ~ 3.7 km
(12,000 ft) AGL in these images and is associated with a tornado near Oklahoma City.

4-62
4.7.4.4.1 Recognition of a Tornado Vortex Signature. In addition to the
appearance of a hook-shaped echo in a low level Reflectivity product, the mesocyclone detection
capability of the WSR-88D offers clues to the existence of tornadic circulations and possible
tornadoes.
Unlike the initial version of the TVS algorithm which would confine searches to only the
mesocyclone vicinity, the TDA (Part C, Chapter 3, of this Handbook) will search the velocity
data for a TVS signature. If identified, the TVS product provides information on potential
tornadic circulations and the location of the signatures. The graphic overlay product can be
used on a low-level V or R product, or the SRR, SRM, or SWA products in order to help
identify the rotational phenomenon associated with the detection of the signature and to verify
its existence and location. Again, base products must always be used to verify algorithm
output, whether negative or positive.
Monitoring V, DV, and SRR, SRM, and SWA products can be useful in identifying
mesocyclone and TVS existence and development. TVS rotation in a storm may begin in mid-
levels aloft, co-located with updraft and often centered within or near the BWER, then extend
downward as well as upward with time. However, these signatures are often the “non-
descending” variety and develop near the surface within a strongly convergent flow and/or
within or beneath a mesocyclone.
Trapp et al. (1999) examined a diverse data set, comprised of 52 events varying from southern
Great Plains supercell tornadoes to tropical cyclone-spawned tornadoes. They stated that when
classifying each by an objective means, they found (with a standard error of 7%) that 52% of
the sampled tornadoes had descending TVSs, and 48% had non-descending TVSs. The results
were stratified according to attributes of the tornado and TVS. For example, the descending
TVSs were associated with greater differential velocity and greater tornado lead time.
Tornadoes within squall lines and bow echoes tended to be associated with non-descending
TVSs, an identification which provided a mean tornado lead time of 5 min. Based on these
results, they concluded that radar operators should recognize, while interpreting Doppler radar
signatures for tornado warning decisions, that a large percentage of tornadoes form in a manner
different than Brown et al. (1978) who found that TVSs consistently descended. But then
Brown et al. (1978) had a far smaller data set.
Tornado Vortex Signature characteristics, including the descent or non-descent, can be

monitored via displays of different elevation angle velocity products in multi-panel displays.
Inadequate removal of storm motion may somewhat obscure these phenomena, but this is
usually minor since only a constant is added to the velocity field.
4.7.4.4.2 Considerations. The TDA is actually a misnomer. Rather than

identifying only tornadoes, the algorithm detects a variety of vortices, including tornadic
vortices. Thus, it must be used with caution for tornado detection.

4-63
The TDA in its current form uses a variety of adaptable parameter data sets to tune the
algorithm to various conditions such as tropical cyclones, squall lines, or large Great Plains
thunderstorms. These data sets better adapt the algorithm functionality to changeable
conditions and improve its performance. However, even with these improvements, the TDA
still has a rather high false alarm rate. Additionally the algorithm detects both surface-based
(lowest radar sampling level) TVSs and Elevated TVSs (ETVSs). In most cases, the ETVS
portion of the algorithm is disabled because those TVSs are not often correlated with tornadoes
and may be confusing to the user. F0 and F1 tornadoes (i.e., gust front tornadoes, those
developing from a bow echo, non-supercell tornadoes, and along convergence boundaries) may
not have a parent mesocyclone. Tornadoes at higher altitudes (in mountainous areas) are often
small and may lack well-defined signatures making recognition difficult.
Due to beam broadening, detection of a TVS is limited to a range of about 55 nm. Because of
beam broadening, beyond that range, strong gate-to-gate shears are typically strong mesocyclones
instead of actual TVSs. Thus, additional information regarding the increased likelihood of
tornadoes gained through the TVS recognition is absent at longer ranges. However, the relative
strength of the mesocyclone may provide a clue to a tornado's existence or likelihood, even at
longer ranges.
Typically, the TVS and even the mesocyclone are detectable within SW products as broadened
widths. But these broad widths cannot be substituted as recognition criteria for tornadoes because
there are many other causes of broad SW. In addition, at longer ranges, there is a general increase
in SW owing to the shear of the horizontal winds with height and the broad beamwidth. In
addition, the mesocyclone itself is typified by broad spectrum widths across the signature center.
A tornadic signature may be located at the periphery of the storm echo. This is especially true in
low-levels on the edge of the storm echo itself or at the edge of the hook echo where weak power
returns result in a low signal-to-noise ratio (where sidelobe contamination may also occur), thus
hindering signature recognition. Portions of the supercell WER and BWER are also areas of weak
signal return and the area where TVSs occasionally develop, making recognition of tornadic
signatures difficult in these areas, as well.
The greatest potential for tornado development is indicated when a tightening/deepening

mesocyclone circulation is noted on the SRM, a TVS is identified (operator), strong low-level
convergence below a mesocyclone base is detected, a storm moving into a low LCL and/or strong
low-level shear environment is seen, and when a storm is interacting with a boundary.
We have addressed here the TVS associated with the supercell because of the signatures overriding
importance and because a TVS is usually associated with stronger or violent tornadoes. Weak
tornadoes, those from F0 to, perhaps, F2 on the Fujita scale, are more often not associated with
detectable TVSs. However, TVSs, manually or algorithm detected, also do occur with non-
supercell storms, although rarely.

4-64
4.8 Hail. Thunderstorms with strong and extensive updrafts, high super-cooled liquid water
content, and large, supercooled cloud drop sizes are favorable for hail formation. Essentially, the
features important to large hail formation are: strong storm-relative flow, large CAPE in the hail
growth region of the atmosphere (-20 to -40 deg C), strong, deep layer shear (> 15.4 ms-1 (30 kts)
in the 0-6 km layer), a strong mesocyclone, a storm WER, a BWER, high reflectivity, and high
VILs (and VIL density). The resulting hail can be spherical or irregular in shape and sometimes
extremely large.
4.8.1 Recognition of Hail Potential. Historically, one of the first techniques used to
identify storms producing hail was to identify storms that possessed high reflectivity > ~ 55 dBZ.
Subsequently, storms that possessed unusually high reflectivity at unusually high altitudes within
the storm were found to produce hail, often large in size. An intense core in a hailstorm will
usually begin developing at higher elevations in the storm and then descend toward the base of the
storm cloud. Therefore, a quarter-screen display of successively higher heights may enable the
detection of the descent. An RCS product accurately placed through the storm can aid in the
identification of strong mid- or high-level reflectivity magnitude. In addition, the LRM product
can be used to isolate those storms with abnormally high altitude and high magnitude reflectivity
and likelihood for large hail. However, while LRM layer-heights are established, the altitude
thresholds for this high reflectivity are variable and change with season and with airmass.
High values of VIL (Green and Clark 1972) and the related severe weather probability values have
been used in the past as estimators of hail potential, as well as other forms of severe weather. In
fact, VIL was one of the more reliable products. However, we also know that the critical VIL
values signifying hail vary with season, with airmass, over the radar coverage area, and even as a
function of the Volume Coverage Pattern. Additionally, a “critical VIL value” is normally needed
to establish the threshold for hail, given an airmass or particular day. Therefore, at this time, use of
VIL as a hail discriminator is not highly recommended although it can be somewhat useful.
A second related parameter has been developed, that of VIL Density (Amburn and Wolfe 1997;
Cerniglia and Snyder 2002; Blaes et al. 1998). VIL Density is VIL normalized to storm depth.
This was done in part to accommodate VIL values that are smaller simply because the storms in
question are mini- or “moderate” supercells which are still very significant from a severe weather
perspective, yet will have smaller VIL values by reason of their smaller stature. While certain
values are indicative of hail, estimating hail size through the use of VIL Density shows very little
skill according to Edwards and Thompson (1998).
The structure of convective storms in a sheared environment has been found to provide an excellent
clue to the potential for intense updraft and large hail with a storm. For example, storms exhibiting
an overhang, related WER, and maximum storm top located above the low-level, tight reflectivity
gradient (or the WER itself) have been found to be large hail producers (Lemon 1978; Burgess and
Lemon 1990). Generation of these same three-dimensional echo features is related to the
abnormally large vertical mass transport and the intense divergence at updraft summit. In fact,
Witt and Nelson (1991) have documented the direct relationship of hailstone size to divergence

4-65
strength (Figure 4-32).
Note that storms exhibiting this three-dimensional reflectivity structure are often supercells, as
well. These same storms containing mesocyclones are often prolific producers of very large hail (>
5 cm (2 in) in diameter). Mesocyclonic storms seem to be associated with airflow (and liquid
water concentrations) that favor long hailstone residence-times within the storm where very rapid
wet hail growth occurs. Thus, detection of mesocyclones via V and DV products and, therefore,
supercells, are strong indications of large or very large hail.
We now have an improved HI that uses the HDA output as explained in Part C, Chapter 3, of this
Handbook (Witt et al. 1998). The algorithm output includes the POH and POSH, as well as MEHS
with the storm cell in question. This algorithm, dependent on the current thermodynamic
properties of the atmosphere within the County Warning Area (CWA) of the associated Weather
Forecast Office and, as determined by a local, recent, atmospheric sounding, bases its output on the
distribution of reflectivity relative to the environmental thermodynamic profiles. While subject to
algorithm limitations, this algorithm has proven to function far better than the previous hail
algorithm it replaces.
Products that can be used for large hail identification include the reflectivity and velocity products
and the high resolution versions, displayed using the multi-panel methods with R, V, and DV. In
addition, properly placed Vertical Cross Sections (RCS and VCS), as well as LRMs, VIL, and the
output of the hail algorithm and product, HI (updated with the current sounding information) are
effective products for hail detection.
Although divergence measurements are not provided directly by the WSR-88D, a VCS through the
storm may provide information on the magnitude of the divergence at the storm summit, which
may help assess the potential for hail. Vertical Cross Section placement is critical. However, use
of high elevation angle velocity products may be superior to attempted use of the VCS product.
The limitation of velocity product usage is that both the maximum outbound and inbound velocities
may not be present in a single elevation angle velocity (or SRM and SWA) product.
Finally, the Three-Body Scatter Spike (TBSS) as occurs in the Mie scattering range of radars can
be used (Figures 4-33 and 4-34). Use of this operational hail identification and related warning
criteria was proposed by Lemon (1998). Lemon’s work was based on Zrnic (1987) and Wilson and
Reum (1986, 1988).
For operational application using the WSR-88D, Lemon (1998) recommended interpreting the
signature as indicating hail greater than 2.5 cm in diameter is falling or will fall with the storm.
The TBSS signature for large hail identification is a sufficient, but not necessary, condition.
Further, because storms associated with the TBSS are also often high-end damaging wind storms,
the user should anticipate the potential of winds up to 41 ms-1 (80 kts) or more with the signature-
bearing storms.

4-66
Figure 4-32
Storm Summit Divergence vs. Hailstone Size
Maximum hailstone diameter (vertical axis) as a function of the maximum measured storm
summit divergence values (horizontal axis). Note the nearly direct relationship between
the two quantities. From Witt and Nelson (1991).

4-67
Figure 4-33
Radar Signal Path Schematic
Schematic of the radar signal path responsible for the TBSS (or flare echo). The dark
shading near point C represents the 60+ dBZ core responsible for producing the artifact.
From Wilson and Reum (1988).

4-68
Figure 4-34
Three-Body Scatter Spike
Oklahoma City, OK WSR-88D 4-panel Reflectivity products at 22:23 UTC on 18 June 1992
(PUP display). The images were made at the same time a hailstorm was bearing down on
Cashion, OK (bottom center of each panel). Images are magnified eight times; data
resolution is 1 km (0.54 nm). Radar data elevation angles are: upper left, 0.5°; upper right,
2.4°; lower left, 3.4°; and lower right, 6.0°. The white cross is located on the reflectivity
core responsible for TBSS and the cross section axis extends through the TBSS to the rear
of the storm in 2.4°, 3.4°, and 6.0°. From Lemon (1998).

4-69
4.8.2 Considerations. High, near-surface reflectivity values (over ~60 dBZe) often
indicate that the precipitation is in the form of hail and not rain. The existence of sidelobe
contamination may also indicate the presence of hail.
Within about 93 km (50 nm), VCP 11 and 12 will provide better coverage for derived product
estimates (such as VIL) and cross sections through storms than will VCP 21.
A combination of products should be used to assess the potential for hail, i.e., RCS, and VCS, VIL,
LRM, the HI product, and R, V, and DV (especially for mesocyclone detection and storm summit
divergence) products. Note that normally the radar is configured to measure velocities with 0.5 m
s-1 resolution. This limits the velocity resolution of the system to +/- 62 ms-1 or +/- 123 kts.
However, because of the intense storm summit divergence within very large hail producing storms,
the velocity resolution of the system should be changed to 1 ms-1 before these measurements are
made. This increases these system velocity measurement limits by a factor of 2.
In order to process the necessary information, the Hail algorithm needs to sample the full volume
and depth of the storm. This seriously limits the HDA output when the storm is located within the
radar “cone of silence.”
As a review, for ordinary storms, the user will mostly be evaluating updraft strength. They will
need to determine updraft location, via the mid-level reflectivity core. Hydrometeor growth is
most rapid as these hydrometeors pass through the most intense part of the updraft and through the
-20° C to about -40° C layer. The intensity and elevation of the core increase as the updraft
intensity increases. So, for severe hail (and potential damaging winds and microbursts), the
maximum heights of the 45, 50, or 55 dBZ reflectivity core (this may vary by season and airmass)
should be monitored. For the vast majority of pulse storms, a warning forecaster must infer the
presence of large hail using a variety of techniques including a higher than normal elevated
reflectivity core, high VIL and VIL density, the presence of low-level reflectivity > 60 dBZ, and
high values of the POSH. This does not include large hail signatures for storms within a
moderately to strongly sheared environment associated with supercells including a Mesocyclone, a
WER, and a BWER. There is a direct observation of large hail within the beam when a WSR-88D
or S-band weather radar detects the TBSS signature.
The types of radar products useful for analyzing potential hail-producing storms are: R/SRM
combinations using multi-panel displays, high resolution Reflectivity and Velocity products
and LRM, CR, VIL, and HI products. In some situations, the user can evaluate hail potential at
the same time they are evaluating its tornado threat as some of the features looked for are the
same (e.g., WER, BWER, a strong mesocyclone) for both threats. Time-lapse loops using 6 to
10 frames can be useful to monitor motion and development.
4.9 Extratropical Cyclones. In mid-latitudes, upper level, long wave troughs, and short-waves
generate low pressure systems with a characteristically well-defined region of warm air advection

4-70
ahead of the disturbance and cold air advection to the rear. Extratropical cyclones (ETCs) typically
have a diameter exceeding 500 km (270 nm) which, if centered on a radar, more than encompasses
the area of radar coverage. This means that, most of the time, only a partial picture of an
extratropical cyclone is obtained from a single radar. Thus, use of regional or national mosaics
generated from the radar network as a whole are normally essential in order to obtain the full
picture of an extratropical system (Figures 4-35 and 4-36). However, at times, a single radar may
detect a circulation pattern associated with the cyclone.
4.9.1 Evolution of Stratiform Clouds and Precipitation. The evolution of cloud and
precipitation layers can be monitored as described previously in Section 4.5. In addition, the VWP
is useful in determining the formation of stratus clouds, as well as monitoring advection patterns.
For example, with the derived wind information and use of upper air charts the user can monitor
moisture and thermal advection with time. Moreover, the VWP can be used to evaluate numerical
model performance.
The ET can be used to aid in monitoring the development of precipitation over time. Increased
returns and identification of cloud layers and cloud tops in Reflectivity data (best in Clear Air
Mode) can be expected. Otherwise, the user can monitor trends in precipitation development
in mid-levels via mid-level Reflectivity products.
4.9.1.1 Recognition of Precipitation Trends. A time lapse sequence of

Reflectivity and Precipitation Accumulation products, as well as the CR products (Figures 4-35
and 4-36), can be used to provide information on the extent of coverage, intensity, and trend of
the system and associated precipitation. The local radar CR provides trend information via
changes in maximum reflectivity throughout the volume, and the associated Combined
Attribute Table provides output from the majority of algorithms. Of course, some of the
reflectivity in the CR products may be thunderstorm anvils or precipitation aloft not yet
reaching the surface (virga). Reflectivity Cross Section products can also be helpful in
determining the depth of a layer of precipitation and the change in this layer over time.
Much of the precipitation in these ETC systems is stratiform precipitation characterized by growth
through collisions and coalescence. Precipitation of that character is most intense at or very near
the surface and, therefore, reflectivity is strongest at the lowest levels scanned. Because of the
improved low-level scanning provided by VCP 12, this is the preferred VCP in extratropical
systems. Reflectivity data is used in much the same way as it is used in convective storm events
(i.e., the base data is used primarily in 4-panel reflectivity and velocity mode). This is done to
obtain the best four-dimensional picture of the evolving precipitation areas, and embedded heavier
precipitation bands caused by frontogenesis, isentropic lift, gravity waves, and/or slantwise
convection (Shields et al. 1991; Moore and Kaster 1993).

4-71
Figure 4-35
National Mosaic of Base Reflectivity of an Extratropical Cyclone
A national mosaic of Composite Reflectivity products obtained from WSR-88D radars at

23:30 UTC on 5 April 2005. The precipitation shield and convection centered over Kansas
was created by an extratropical cyclone. The cyclone with a deep upper low pressure and
surface system is centered over southwest Kansas. A frontal line of thunderstorms is over
east Oklahoma and northeast Texas while a more stratiform area of rain is centered over
western Kansas. (Courtesy, National Severe Storms Laboratory).

4-72
Figure 4-36
Regional Mosaic of Base Reflectivity of an Extratropical Cyclone
Same as Figure 4-35, except as a regional view of the central Great Plains. (Courtesy,
National Severe Storms Laboratory).

4-73
4.9.1.2 Considerations. The ET is limited due to its reflectivity threshold (18
dBZe) in its ability to effectively monitor precipitation development and trend. This threshold will
need to be lowered for stratiform and other events. Determining necessary changes to threshold
levels may require on-site directives and operational experience.
During stratiform events, since the main concern is usually the weather in the lower two or
three elevation angles, operating the radar using VCP 12 or 121 will often provide adequate
vertical resolution. Moreover, the extent of range-overlaid echo can be significantly mitigated
through the use of VCP 121. This is especially important because of the likelihood of
widespread precipitation and range-overlaid echo.
4.9.2 System Movement. Within range of the radar, a closed circulation around the center
of an ETC is normally not observed unless, perhaps, the system is deep and intense. Normally,
only a portion of the circulation is evident. Rather, only a portion of the wind field can be
observed by individual radars. However, cyclone-associated low-level jets are often observed in
the radar-centered velocity products with modest elevation angles. The VWP may be used in
conjunction with these products as well. Moreover, in the velocity and Reflectivity products, fronts
and other discontinuities associated with these circulations are often detectable as they near
individual radars. This has allowed much more accurate frontal placement than in the past.
4.9.2.1 System Movement Recognition. Using individual rain shower or

thunderstorm cell motion to determine overall cyclone movement is not recommended since cell
movement does not match total system movement. The SCIT and its STI product should also not
be used in attempting to determine (MCS/Mesoscale Convective Complex (MCC)) motion within
the overall ETC. The algorithms will track individual cells, but these storms move with the local
wind-field -- not with the larger-scale system.
Unless the ETC center is strong and small, a single radar cannot observe its center. Rather, a
mosaic of the lowest level reflectivity data is needed and used with time-lapse. Sometimes the
CR product can better portray circulation about the storm center. But even when using mosaic
products, a closed circulation is rarely seen. Somewhat smaller scale convective systems
within the ETC have resolvable vortex centers. The Mesoscale Convective Vortices (MCVs)
are often associated with MCSs which, in turn, may be associated with the ETC. These MCVs
are detected within mid-levels and the residual mid-level cloud, virga, and rainfall. Within the
context of the extratropical cyclone, centers of these MCV systems and their tracks are often
critical for later thunderstorm and MCC re-initiation. The feature tracking capability can be
used for tracking these MCVs.
4.9.2.2 Considerations. Often, synoptic systems in their more mature stages

of development are characterized by comma-shaped cloud and precipitation systems.
Typically, there is not sufficient return from all sectors of the cyclone to detect and track the
complete circulation. However, when there is widespread precipitation along the cold front,
extensive warm-frontal precipitation, and precipitation within the cloud head, then it is

4-74
possible that the entire cyclone system (with the exception of the dryer air sector to the rear)
may be tracked via the radar network mosaics. Thus, the network reflectivity mosaic
capability provides an excellent comparison with numerical model forecasts.
4.9.3 Cell Movement. A characteristic feature of ETCs is the presence of narrow

convective precipitation bands in both the warm and cold sectors of the cyclone. Within these
bands, embedded cells with high values of reflectivity are not uncommon.
4.9.3.1 Cell Movement Recognition. Embedded cells can be easily identified

and tracked by the SCIT algorithm from any individual radar. Generally, these cells move faster
than the precipitation bands and to the left of the mean band motion.
4.9.3.2 Considerations. In cold ETCs, reflectivity values of individual cells

rarely exceed about 45 dBZe. Higher reflectivity values, sometimes ranging from 50 dBZe to 60
dBZe, in precipitation bands, may be the result of large, wet, snowflakes and/or ice pellets within
the melting level. Much of the reflectivity remains at or below 30 dBZe, but these will probably
be tracked by the SCIT algorithm. Precipitation bands, like convective squall lines, are generally
not tracked by the SCIT algorithm since it isolates individual reflectivity cores. Therefore, a
time lapse of the Reflectivity product should be used to deduce overall complex or line motion.
4.10 Winter Storms. The WSR-88D provides data useful in monitoring the increase and
lowering of ice crystal-laden clouds, in estimating precipitation onset at the surface, the height of
the melting level, in tracking and predicting the motion of precipitation regions, mesolows, high
wind areas, and heavy snow. Snow is generally more effectively detected by the WSR-88D, in
contrast to past radars, but snow will be more difficult to identify than liquid precipitation. The
dielectric constant of water is nearly 1.0, while that of snow is only ~0.20; thus, snow is a much
less efficient reflector of microwave radar energy. Additionally, the small particle size and a
more complex pattern and geometry of snowflakes further reduce microwave backscattering.
4.10.1 Onset of Surface Precipitation. With the WSR-88D, it is possible to monitor the
development, advancement, and/or presence of cloud, inversions, or descending or evaporating
precipitation layers, and infer height estimates from cursor readout information.
Cloud layers, inversions, and virga result in radar echo (reflectivity, velocity, and spectrum
width) signatures best described as radar-centered "donuts" (Figures 4-10 and 4-37).
Considerable information, produced from a variety of meteorological situations, can be derived
from these signatures (Lemon and Quoetone 1994). Examples include base and top heights of
cloud layers, slope of a layer, estimation of precipitation onset, freezing level heights and descent
rates, timing of precipitation phase change, and perhaps even strength of thermal gradients. The
VWP is also a very useful tool for application to these same storms and situations. As mentioned
earlier, it is possible to monitor the winds aloft and to compare them to numerical guidance. It is
also possible to monitor the base of virga aloft and its rate of descent (Figure 4-37).

4-75
Figure 4-37
Precipitation Mode Reflectivity Scan of Virga
Oklahoma City, OK WSR-88D Reflectivity product at 18:45 UTC on 12 December 2000

(AWIPS display). This is a radar-centered “donut” echo created by virga aloft, with the
inner edge of the donut at the base of the echo and the outer edge being the top of the
detectable return of the precipitation. The surface observations are overlaid.

4-76
It is possible to estimate the onset of precipitation as it develops toward the surface. For example,
the radar is located north and east of an eastward-moving mid-latitude cyclone with the associated
frontal system south of the area. Commonly, these storms begin by a gradual increase in high and
mid-level cloudiness, the formation of multiple cloud layers that advance progressively from the
southwest or west, and a corresponding lowering in cloud bases. "Overrunning" has often been
used to describe these events, owing to the presence of a surface warm front upstream (relative to
mid-tropospheric flow). Echo monitoring can be done using R and V products in a multi-panel
display and in time-lapse mode. Echo Tops can also be used in this process but again, the
reflectivity threshold is ~ 18 dBZe. Reflectivity aloft near cloud and precipitation tops is often low,
ranging from -10 to +15 dBZ early in the process. Especially in higher latitudes during the fall and
winter months, precipitation aloft gradually develops from these advancing cloud layers but falls
into dryer air and evaporates. This process both cools and moistens the troposphere at
progressively lower levels until rain or snowfall reaches the surface. This, in turn, is often well
observed by the WSR-88D. In the absence of algorithms, the operator can manually estimate the
base descent rate and arrive at an estimated time of the onset of surface precipitation at the radar
and other locations. The base of the descending precipitation will usually be lower toward the
direction from which the clouds advance. The freezing level, if aloft, will also often slope in the
same sense.
4.10.2 Cold Air Depth, Freezing Level. When, within the precipitation-filled volume,
precipitation changes from snow to rain during its descent, and convection is neither vigorous
nor widespread, there is also frequently a radar detected "bright band" (Figure 4-39). This
relatively highly reflective freezing/melting layer is known as the bright band from observations
made using range-height scanning and display early in weather radar history (Battan 1959). As
in the detection of inversions, cloud layers, or precipitation aloft, and owing to the geometry of
using an elevated scanning beam, this confined phase-transition layer is detectable as a ring of
enhanced reflectivity of finite depth centered about the RDA (Lemon and Quoetone, 1994).
As explained, the melting level of uniform height can be identified as a ring, or partial ring, of
higher reflectivity values in the Reflectivity product (Figure 4-39). If the height of the melting
level varies or slopes within the range of the radar, the ring will be asymmetric (closer to the
radar on one side than on the other), or the bright band may appear as an arc or as an irregular
band. The signature will coincide with one of slightly higher spectrum width values (perhaps of
~8 knots). The increase or decrease in altitude of this feature can be monitored.
Reflectivity Cross Section products can be useful in identifying the melting level. An exact base
and top cannot be obtained due to the variance within the melting level of frozen versus liquid
water droplets, distance of the phenomenon from the radar, thickness of the layer itself, and the
ambiguity of radar height measurements due to beam width and refraction index variation.

4-77
Figure 4-38
VAD Wind Profile Product
Oklahoma City, OK WSR-88D Velocity Azimuth Display (VAD) Wind Profile product at
18:03 UTC on 12 December 2000 (AWIPS display). The base of the virga aloft is about
7,000 ft and the wind profile shows west-southwest winds above the low-level easterlies in
the colder air below the frontal inversion.

4-78
Figure 4-39
Bright Band
Springfield, MO WSR-88D Reflectivity product at 11:28 UTC on 29 January 2001 (AWIPS

display). This product, 2.4° elevation angle, shows a bright band (dark red ring of 40 dBZ
to 50 dBZ) surrounding the radar.

4-79
The observation of the band is useful because it denotes the environmental freezing level and a
region of precipitation phase transition from frozen to liquid. It is also usually a region of abundant
cloud water. Cursor-indicated height, when placed on the outer ring edge, suggests the transition
height from dry snowflakes to moist snowflakes. (The dendritic snow crystal growth zone is above
this region where temperatures are nearer -12° to -15° C). Inner ring height locates the transition
altitude from wet snow to rain.
There is also the ability to monitor the critical depth of the cold air as is illustrated in Figure 4-38.
In this case the cold air (with easterly winds) appears to be very shallow, although from 1.2 km to
2.1 km (4,000 ft to 7,000 feet) there is no data displayed. This, in part, may be due to the lack of
tracers, yet there is also the likelihood of asymmetric flow due to the frontal surface itself. Cold air
depth will often dictate if the precipitation is liquid or solid when it reaches the surface.
Through the use of multiple radars and/or use of a single radar over time, the slope of the bright
band (and the freezing level) can be mapped. Identification of the freezing level and areas of
mixed precipitation will permit specificity in nowcasting and short term forecasting. Additionally,
within mountainous areas it is possible to accurately estimate the rain and snow levels and areas of
mixed precipitation. At times, a time-lapse loop may be needed to reveal a bright band rather than
passing shower cells if it is present.
4.10.3 Snowfall Estimation. Algorithms have been added to the WSR-88D that estimate
both snowfall accumulation (Snow Depth, SD) and liquid water equivalent (Snow Water
Equivalent, SWE) with that snowfall. Six new products are used. Those products are very similar
to the rainfall products because they cover the same time-periods. When the user is within the
snowfall portion of the storm, north of the surface low track, these products are very useful. In
order to estimate snowfall and the water equivalent over a one-hour running period, the user
requests the One-Hour Snow Depth (OSD) product (Figure 4-40).
The OSD is used on a continuous basis during significant snowfall. In order to maintain a record
of the water equivalent of this snowfall, the Snow Water Equivalent (OSW) is requested (Figure
4-41).
When snowfall accumulations are expected to occur over an extended period and to be
significant, then the Storm-Total Snow Depth (SSD) product can be requested (Figure 4-42).
Like the OSW, there is also the water equivalent product (not shown), the Storm Total Snow
Water Equivalent (SSW) product. This measure of snow water increases in importance as the
weather warms after an extended period of snow and rains fall on significant snow packs with
considerable melting. Floods frequently result. Thus, the amount of stored snow water becomes
very important in estimating potential runoff and potential flooding, sometimes the sole source of
runoff. Of course, periods of sublimation may deplete the snow water significantly.

4-80
Figure 4-40
One-Hour Snow Depth Accumulation Product
Norman, OK WSR-88D (test bed radar) One-Hour Snow Depth Accumulation product
(OSD) at 11:32 UTC on 7 February 1999 (CODEview graphic). The radar-estimated snow
accumulation, snow depth that occurred over the last one hour period is included.

4-81
Figure 4-41
One-Hour Snow Water Equivalent Accumulation Product
Norman, OK WSR-88D (test bed radar) One-Hour Snow Water Equivalent Accumulation
product (OSW) at 11:32 UTC on 7 February 1999 (CODEview graphic). The liquid water
content of Figure 4-40 is shown here.

4-82
Figure 4-42
Storm Total Snow Depth Accumulation Product
Norman, OK WSR-88D (test bed radar) Storm Total Snow Depth Accumulation product
(SSD) at 11:32 UTC on 7 February 1999 (CODEview graphic).

4-83
At times, the user wants to select a specific period of snowfall accumulation such as from 0000
UTC to 2400 UTC. The User Selectable Snow Accumulation (Depth) product, USD, is used for
such a purpose (Figure 4-43). The accompanying product used to estimate the snow water
equivalent for a user selectable interval is the User Selectable Snow Water Equivalent (USW)
product (Figure 4-44).
Mesoscale snowbands, perhaps only a few miles in width, but producing heavy snow, are
common in winter cyclones (Figure 4-45). These bands are typically 90 km to 160 km (50 to
100 nm) to the left of the surface low's track. Intense wind shears and strong frontal zones, also
found in these mid-latitude winter storms, often result in significant mesoscale vertical motions
and corresponding slantwise convection (sometimes accompanied by lightning). These snowfall
products will be quite helpful for monitoring the type of snowfall.
4.10.4 Recognition of Heavy Snow and Snow Showers. During snowfall situations, it is
not unusual to see narrow bands of higher reflectivity values in a Reflectivity product that often
indicate areas of heavier snowfall. This banding, often related to slantwise convection, may also be
identified occasionally in the velocity field. The 4-panel or multi-panel Reflectivity products are
the best way to monitor these bands.
In the case of both sleet and rain occurring within the sampling area of the radar, embedded areas
of higher reflectivity will often appear (40 to 50 dBZe may be common) as a result. A linear
discontinuity in reflectivity returns will sometimes be seen along the rain/snow line. Spectrum
widths will again be enhanced along this line.
Often found embedded within areas of snow are slantwise convective cells producing moderate to
heavy snowfall at the surface. Sometimes these heavy convective snow regions will be associated
with lightning. These cells generally move faster and to the left of the mean band motion. With
reflectivity values often exceeding 40 dBZe in these convective cells, they will be tracked well
unless segments from adjacent cells overlap.
4.10.4.1 Considerations. The OSD and the USD can support nowcasts by
quantifying snow rates, highlighting areas of higher and lower rates, and showing movement and
trends of these areas. The SSD and longer-term USD products can support updates to forecasts of
total accumulations and verification of snow amounts, especially where there are gaps in snow
spotter coverage. Time lapse loops of these snowfall and water equivalent products will also better
highlight motion and tracks of greater accumulation.
These products suffer less from blowing and drifting snow (most of the blowing snow is below
radar horizon) that plagues spotter surface measurements. However, below beam affects such as
low boundary layer humidity, can lead to overestimation of snowfall.

4-84
Figure 4-43
User Selectable Snow Depth Accumulation Product
Norman, OK WSR-88D (test bed radar) User Selectable Snow Depth Accumulation
product (USD) at 12:13 UTC on 7 February 1999 (CODEview graphic).

4-85
Figure 4-44
User Selectable Snow Water Equivalent Accumulation Product
Norman, OK WSR-88D (test bed radar) User Selectable Snow Water Equivalent
Accumulation product (USW) at 12:13 UTC on 7 February 1999 (CODEview graphic).

4-86
Figure 4-45
Convective Snowfall Reflectivity
Boston, MA WSR-88D Reflectivity product at 15:01 UTC on 23 December 1997 (legacy

PUP display). The embedded heavy (yellows) and very heavy (red) areas depict
convective snow.

4-87
Moreover, variable Z-S (Reflectivity – Snowfall rate) relationships will also significantly
contribute to snowfall and water equivalent inaccuracies. For example, air temperature contributes
to the amount of liquid water that is contained in the snow, and more importantly, its reflectivity.
For comparable snowfall rates the reflectivity can vary appreciably. Thus, one of the most
significant limiting factors in snowfall estimation is the Z-S relationship.
User selectable products can be requested to coincide with other regular operational products (e.g.
24-hour totals to coincide with daily climate summaries). Reliable snow measurements (such as
trained spotter reports), if available, should be used in conjunction with Snow Accumulation
Algorithm (SAA) products to validate and calibrate the algorithm estimates. However, if no
reliable and timely snow measurements are available, WSR-88D snow products may be the only or
best source of snow estimates. For example, during severe snow storms with strong winds,
spotters may be unable to make accurate, or even regular, measurements and report them in a
timely manner, or spotters may not be directly under the echoes of interest (such as the heaviest
snow bands).
The SAA products will have limited range, perhaps out to about 75 to 90 km (40 to 50 nm) owing
to overshooting the dendritic snowflake formation zone and overshooting the convective cloud
layer. This is the result of the relatively shallow cloud layers characterizing many snow storms.
This will especially be the case with the very shallow lake-effect storms. Other complicating
features include contamination due to the bright band and evaporation below the radar horizon
which can’t be compensated for. Additionally, in the cases of strong low-level winds, there is the
problem of horizontal drift of falling snow. Depending on conditions and radar sampling depths,
falling snow may be carried up to a few kilometers before reaching the surface while below the
radar horizon. Finally, irregular precipitation particle shapes will not fit well with the Z-S
fundamental assumptions.
The above-mentioned lake-effect snow storms occur in the form of shallow convective bands that
nearly parallel low level winds. At times, they may be only a few thousand feet in depth, yet can
produce extremely heavy snow with near-zero visibilities. Occasionally, strong winds and
lightning will also accompany these occurrences. Sampling these events using the Clear Air scan
strategies is often effective as is the case with many non-lake effect snow situations. However,
owing to the shallow nature of these events, radar is effective at times only to about 35 to 50 km
ranges.
Movement of snow bands may be deduced by a time lapse of the Reflectivity and the SSD, the
storm total snowfall product. The SCIT algorithm may not provide adequate information if these
snow bands lack a cellular nature and overlap. Lake-effect snow bands may remain nearly
stationary over prolonged periods.
Embedded higher reflectivity may be the result of the melting level and not heavier snow, sleet,
or convective cells. On the other hand, they may be convective in nature (Figure 4-45) and may
even be associated with lightning. Surface conditions and surface reports are invaluable in these

4-88
cases. Additionally, examining somewhat higher elevation angles may reveal the bright band
and its location relative to the embedded higher reflectivity and may provide an indication of
their relationship.
4.10.5 Rain/Snow Line. Determining the location of the rain/snow line is often difficult.
The base data products from the WSR-88D may prove somewhat useful in this, but integration of
all types of data are usually needed to locate this change. But, when present, the bright band can
assist, especially when one side of the ring is much closer to the radar than the other, indicating a
sloping freezing level. Additionally, the area of transition in low-level Reflectivity products is
often marked by higher reflectivity values, whether cellular or stratiform, where there are large,
wet snowflakes falling in a mixture with rain, or where rain, sleet, and snow are mixed. In these
regions reflectivity may even exceed 60 dBZe. Additionally, areas of rain are often more cellular
in nature with higher reflectivity gradients. This can help in identifying regions that are only rain
or only snow.
4.10.5.1 Recognition of a Rain/Snow Line. It might be possible to determine

the rain/snow line through use of the Spectrum Width products. These products may provide a
means to help identify temperature discontinuities. Used in conjunction with the Reflectivity
product, Spectrum Width products might be used in certain situations to estimate where, within
the reflectivity field, the rain/ snow line is located. The transition region is often characterized
by somewhat higher SW values. But, there are a variety of reasons for broadened spectrum
widths, so caution must be exercised and SW should never be used alone. Integration of all
observations is needed.
4.10.5.2 Considerations. Cross sections of base data products may provide an

indication of frontal position. However, the VWP is most effective in identifying the frontal
surface aloft above the radar and changes in the cold-air depth. In most situations, surface reports
will be most useful in locating the rain/snow line.
4.11 Tropical Cyclones. In the fully developed stage, a tropical cyclone is a nearly circular,
warm core vortex that extends throughout the depth of the troposphere and has a typical
horizontal scale of hundreds of miles. Maximum horizontal winds occur 10 to 55 km (5 to 30
nm) from the center and 0.5 km to 1 km (1,600 to 3,300 ft) above the surface (Willoughby et al.
1984) (Figure 4-46).
The WSR-88D provides Velocity and Reflectivity products allow the tracking and anticipation of
wind and precipitation patterns associated with tropical cyclones before during and after they
make landfall. As these destructive phenomena come within range of coastal radars, information
on system center, spiral rain bands, and wind distribution will increase. In mature tropical
cyclones, eye-wall diameter and quantitative estimates of intensities are available.

4-89
Figure 4-46
Hurricane Charley Reflectivity
Tampa, FL WSR-88D Reflectivity product at 20:52 UTC on 13 August 2004 (NCDC

NEXRAD viewer). This image was taken as Hurricane Charley made landfall. The eye
diameter is about 11 km (6 nm).

4-90
The effects of the Doppler velocity-range ambiguity (Part B of this Handbook) are especially
critical during the observation of tropical cyclones. The legacy WSR-88D scan strategies (without
VCPs 12 and 121) allow the unambiguous measurement of Doppler velocities out to a range of
only 150 km (81 nm) (Nyquist co-interval of + 25 ms-1 (49 kts). To raise the Nyquist co-interval
to + 31 ms-1 (60 kts), the radar is only able to unambiguously measure Doppler velocities out to
120 km (65 nm).
Thus, tropical cyclones offer the combined problems of extensive precipitation areas with
diameters often exceeding 300 km (160 nm) and wind speeds often exceeding 51 ms-1 (100 kts). In
order to minimize regions of range-folded Doppler velocities, it is necessary to operate with a low
Nyquist velocity. However, the new VCP, 121, greatly aids in reducing range-overlaid data, as
well as simultaneously increasing the measurement of high velocities.
The MPDA, as part of VCP 121, is well suited for tropical cyclone observations. The purpose
for this VCP is to enable a WSR-88D unit to optimize scanning and the volume sampling interval
for changing radar echo patterns while, at the same time, permitting a large decrease in range-
overlaid echo and in incorrectly dealiased velocities. It was designed to mitigate range folding
and improperly dealiased velocities. Thus, this is ideal for the purpose of tropical cyclone
observations. For example, VCP 121 decreases the amount of overlaid echo by 50% to 70% over
that with VCP 21.
It is important for most tropical cyclone-affected weather forecast offices to change the velocity
measurement increment value for the WSR-88D system from the default setting of 0.5 ms-1 to 1.0
ms-1. This allows the system hardware to measure velocities of +/- 124 ms-1 or +/- ~ 248 kts rather
than +/- 122 kts. More information on WSR-88D operations during tropical cyclones can be found
in the National Hurricane Operations Plan; Chapter 7 – Surface Radar Reporting; and Appendix H
– WSR-88D Operations Plan for Tropical Cyclone Events (http://www.ofcm.noaa.gov/).
4.11.1 Embedded Convective Storms and Tornadoes. Precipitation in tropical cyclones

is organized in rain bands that may not completely encircle the storm center. While extensive areas
of stratiform rain exist in a tropical cyclone, embedded areas of deep convection are often present.
Most tropical cyclone-associated tornadoes develop near areas of intense convection in the outer
spiral rain bands, and are generally weaker than those observed in the Midwestern United States.
However, mesocyclones can occasionally be found in the eyewall, and strong tornadoes can, and
do, occur anywhere within tropical systems. As the storm makes landfall, the right-front quadrant
of the storm is the area most prone to these tornadoes.
4.11.1.1 Recognition of Embedded Convection and Tornadoes. As the first

and subsequent spiral rain bands move toward the shoreline, interest will be in locating the
strongest convective activity embedded in the bands, and any mesocyclones and tornadoes
located at the fringes of the bands. The R, VIL, and ET products can provide an indication as to
the intensity of the convection. However, when examining these rain bands for supercells (often

4-91
mini-supercells), the multi-panel high resolution reflectivity (DR), and velocity (DV) products
are most useful. Because these storms can move very rapidly, SRR products for particular rain
bands can used. But this is also made difficult because of the almost continuous change in
direction of motion. (Spratt et al. 1997).
Maximum reflectivity in the eye wall is typically 45 to 50 dBZe (Part B of this Handbook), a lower
value than mid-latitude extratropical thunderstorms. The lower reflectivity is due to smaller drop
sizes and to the lack of hail in tropical cyclones. Outside the eye wall, there are usually several
rain bands that have isolated areas of deep convection and extensive areas of stratiform rain. These
rain bands have fewer vertically-oriented reflectivity cores than the eye wall. However, it is not
unusual for cells in the outer rain bands to exceed the reflectivity of the eye wall itself.
The favored location for tornado development is either the right front or right quadrants of the
storm at landfall or in any quadrant within 25 nm of the center. Small scale (mini) supercells even
with BWERs and hook echoes are detectable using R, V, and SRR products (Figure 4-47). Low-
level convergence, when the radar is near one of these storms, can also be detected. This, too, is
substantially weaker than that typical of extratropical severe convection. As before, multi-panel
base product display is preferred for the lowest 4 elevation angles in order to monitor storm
structure and mesocyclone development and strength. Vertically Integrated Liquid (VIL), ET, and
LRM may also be used to monitor cell significance, but these are of relatively degraded resolution
and value as compared to base products. Time-lapse loops of 6 to 10 frames are useful.
4.11.1.2 Considerations. Tropical cyclone-induced tornadoes typically form

very rapidly, are weak, and have a short life span, much like gust front tornadoes. As seen in
Figure 4-47, these are mini-supercells with small scales in reflectivity and velocities, but with the
supercell features including the mesocyclone. Thus, they are supercells of the mini variety. The
observer should be mindful that these storms and associated mesocyclones can, although rarely,
become quite strong.
4.11.2 Heavy Rainfall. One of the most devastating aspects of tropical systems can be the
very heavy rainfall. Thus, continuous use of the system rainfall products, especially the STP/DSP
is extremely important in these cases (Figure 4-48). Flooding caused by heavy rainfall from a
tropical cyclone making landfall is common. Both extensive areas of stratiform precipitation and
intense rain-bands are responsible for this phenomenon. Climatologically, much of this rainfall
can, at the same time, be very beneficial.

4-92
Figure 4-47
Mini-Supercell and Mesocyclones Associated with Hurricane Ivan
Tallahassee, FL, WSR-88D 4-panel of products at 20:49 UTC on 15 September 2004

(AWIPS display). A mini-supercell associated with Hurricane Ivan is shown. Top left is a
0.5° Base Reflectivity Data Array product, top right is a 0.5° Base Velocity Data Array
product, bottom left is a 1.5° DV, and bottom right is a 2.4° DV. The mesocyclone is very
small and can be seen in the DV products (red and green associated) with the southern
portion of the storm. This was one of several supercells in the outer band of Ivan, which
moved onshore and produced up to F2 tornadoes.

4-93
Figure 4-48
Storm Total Rainfall Accumulation Product of Hurricane Emily
Brownsville, TX WSR-88D Storm Total Rainfall Accumulation product for the period ending
20:30 UTC, 20 July 2005, TX (National Weather Service Radar Web page). The rainfall
was due to Hurricane Emily which made landfall about this time 75 miles south of the
Texas border about 12 hours earlier.

4-94
4.11.2.2 Considerations. Effectiveness of the SCIT algorithm to accurately
identify and track movement of convective cells in tropical cyclones is limited. Because of the
rapid changes in cell character, intersecting storms, or storms in close proximity to each other, their
rapid and curvilinear motion, storm cells, and tracks are frequently reinitialized. However, the
SCIT algorithm, when applied outside of strong tropical cyclones, can be effective in many other
convective situations, such as echo training.
4.11.3 Convective Scale and Straight Line Winds. Common to tropical cyclones making
landfall is widespread damage from strong winds (Figures 4-49, 4-50, and 4-51). The potential for
convective downdraft occurrences with individual cells is limited owing to the nearly saturated
deep layers within the storm environment, but the mesocyclonic winds with mini-supercells are a
threat, as are the locally “straight-line” winds associated with tropical cyclones moving inland.
4.11.3.1 Recognition of Potential Damage from Embedded Convective Cells

and Straight Line Winds. The high-resolution Base Velocity, DV, is the primary product to
assess these velocities and winds because these winds are ground relative and they are the winds
that do the damage. However, Mean Radial Velocity is also useful (Figure 4-49). (But note that
monitoring these velocities is best done when the velocity product color levels have been adjusted
for the stronger velocities.) The SRM and SRR can be attempted in evaluating storm relative flow
and mesocyclone presence and strength but are not appropriate for evaluating ground relative
damaging winds. (However, SRM and SRR are of limited use with the generally curvilinear and
rapid motion of all cells.) The strongest winds are nearly always associated with the higher
reflectivity rain-bands and the eye wall (Figures 4-46 and 4-49). Multi-panel R, DR, V, and DV
products will be most useful in determination of the wind field characteristics and wind speeds.
Through use of time sequenced velocity and Reflectivity products, wind field motion, and changes
can be tracked. This is useful in assessing regions of greatest damaging wind threat from hurricane
winds (Figure 4-50).
An important source for damaging wind enhancement with tropical cyclones that is often
overlooked is the embedded mesocyclone (Spratt et al. 1997). These circulations are often
emphasized as potential tornado producers but can be overlooked as sources for enhanced ground
relative winds. But these mesocyclones do, in fact, pose an enhanced threat for damaging surface
winds above that of the tropical cyclone itself.
For indications of vertical shear associated with surface frictional affects and the typical wind-
field, irregularities and distribution, the multi-panel high-resolution DV and DR products may be
used. These shears are important because it is believed that they contribute to the low-level shear
favorable to the production of mini-supercells and associated tornadoes. But, for the most part,
this friction induced shear must simply be inferred rather than actually observed owing to radar
horizon. The same is true for the actual tornadoes themselves. Tornado Vortex Signatures are
rare with these mini-supercells because most associated tornadoes are small and relatively weak.

4-95
Figure 4-49
Mean Radial Velocity Product of Hurricane Charley
Tampa Bay, FL WSR-88D Mean Radial Velocity product at 20:53 UTC on 13 August 2004
(NCDC graphics). This 0.54 nm resolution at 0.5° elevation product was made as
Hurricane Charley was making landfall. At this time, radar indicated velocities were up to
263 kmhr-1 (142 kts).

4-96
Figure 4-50
WSR-57 Radar Reflectivity of Hurricane Andrew
Miami, FL WSR-57 radar reflectivity eye wall (0.5° radar elevation angle) of Hurricane
Andrew at 08:35 UTC on 24 August 1992 as the storm moved inland. The hurricane is
virtually centered on the shoreline. This is shortly before the WSR-57 was destroyed by
hurricane wind gusts. Source: NOAA Hurricane Research Laboratory.

4-97
Figure 4-51
Surface Observations of Hurricane Andrew
Surface observations as the Hurricane Andrew eye-wall (shown in Figure 4-50) moved
inland near Miami, FL. Source: NOAA Hurricane Research Laboratory.

4-98
Areas of high reflectivity and precipitation rates can be monitored for correlation with the
occurrence of significant winds and damage. This will best be monitored through the use of time
lapse and the multi-panel R, DR, V, and DV products. In addition, use of the VWP product can
be used to monitor velocities and the shear profiles.
4.12 Flash Floods. Flash floods typically result from locally excessive rainfall, nearly stationary
thunderstorm cells, or repeated thunderstorm cells that "train,” passing over the same area. Flash
floods may also be caused by a dam or levee failure, or from a sudden release of water by an ice
jam. In some instances, rainfall may augment rapidly melting snow and ice and cause flash
flooding.
Factors such as storm movement with respect to a river basin's shape and orientation, terrain, and
soil moisture conditions must also be considered when assessing the potential for flash flooding. In
many flash flooding instances it is local hydrological factors that dictate and actually create flood
occurrence. The WSR-88D user must be cognizant of these local affects.
4.12.1 Recognition of a Flash Flood Event. The WSR-88D provides access to several
products that will be helpful in determining the potential for a flash flood event. The most
important products are the Precipitation Accumulation products, especially the OHP and the STP
products (Figures 4-52 and 4-53), as well as Reflectivity products.
Other meteorological dependencies may contribute to flash flooding, such as:
- Storm (system) motion

- Strength of inflow
- Instability of inflow
- Mixing ratio of inflow layer
- Storm-relative flow (mid-upper)
- Seeding efficiencies (line and interstorm)
- Depth of warm cloud layer
These are just some of the meteorological considerations involved in evaluating convective flash
flood threat potential. Thus, combining these parameters with basin characteristics indicate many
of the various considerations outside of radar that must be considered when evaluating flash flood
potential.

4-99
Figure 4-52
One-Hour Rainfall Accumulation Product
Hastings, NE WSR-88D One-Hour Rainfall Accumulation product (OHP) at 00:28 UTC on

23 June 2003 (AWIPS display). The storm was nearly stationary and very severe.

4-100
Figure 4-53
Storm Total Rainfall Accumulation Product
Hastings, NE WSR-88D Storm Total Rainfall Accumulation product (STP) at 00:28 UTC on

4-101
Individual cell motion is often handled well by the STI product. However, evaluating system
motion is much more difficult, being based on factors that control propagation. The mixing ratio in
the inflow layer is one of those and is related to vertical water vapor flux. Storm-relative mid- and
upper-level flow is important for single cell considerations. Can these winds allow recycling of
hydrometeors back into the updraft? Weaker storm relative flow allows a greater chance of
hydrometeor recycling into the updraft which, in turn, increases precipitation efficiency of single
cell events. Interstorm seeding can help increase precipitation efficiencies in those situations
where individual updrafts are inefficient at recycling their own hydrometeors. Look for connected
regions of 30 dBZ or larger reflectivity between individual convective cores. Also, look for
sustained flanking convection moving into, merging, or developing with slow moving storms.
The depth of the warm cloud layer is important. This is the saturated layer that is above freezing
and is the area where collision coalescence of liquid precipitation occurs. If this layer is deep, then
there is a greater hydrometeor residence time for this process to occur and, therefore, result in a
high precipitation rate, even from relatively innocuous convection. Awareness of this process is
most important if convective cells appear deceptively weak, that is, if they have warm cloud tops,
little lightning, and unimpressive reflectivity. If weak convection is occurring in a nearly saturated
environment with weak CAPE and a high freezing level and low LCL, then it is likely that this
convection is characterized by high precipitation efficiencies along with rain drop distributions that
favor the use of the tropical Z-R relationship. See Petersen et al. (1999) for more details on
excessive rainfall from warm clouds.
In addition to the precipitation accumulation products, other useful products include reflectivity
(multi-panels) and cross sections. In addition, VIL, LRM, and ET products can be used to monitor
storm size, strength, and trends. Moreover, the STI products can be useful in recognizing a
developing or ongoing flash flood situation. The location of boundaries and the low-level jet are
very important in creating and sustaining precipitation and flooding. The VWP product can be
very important in detecting and monitoring this process.
4.12.2 Precipitation Accumulation. The Precipitation Accumulation products provide

estimates of precipitation accumulations for OHP, STP/DSP, THP, and USP (Figures 4-52, 4-53, 4-
54, and 4-55). These quantitative precipitation estimates can be compared with Flash Flood
Guidance values provided by the River Forecast Centers (RFCs) to determine flash flood potential.
Time lapse sequences of the OHP, STP, and USP will prove to be extremely valuable in some
situations. The time lapse sequences will aid in identifying stationary storm complexes or a train
of cells with intense precipitation. A time lapse will also aid in the determination of precipitation
patterns within a basin and, potentially, a basin's response.
Used in conjunction with one another, the Precipitation Accumulation products provide
information about the potential for flooding. For example, the THP (Figure 4-54) used with a
later OHP can help in the identification of areas where basin saturation has occurred.

4-102
Figure 4-54
Three-Hour Rainfall Accumulation Product
Hastings, NE WSR-88D Three-Hour Rainfall Accumulation product (THP) at 03:13 UTC on


4-103
Figure 4-55
User Selectable Rainfall Accumulation Product
Same as Figure 4-52, except it is a User Selectable Rainfall Accumulation product (USP)
and only a 2-hour duration beginning at 23:00 UTC on 22 June 2003 and ending at 01:00
UTC on 23 June 2003 (AWIPS display).

4-104
In severe convection-related rainfall estimation, one note of caution is worthwhile. In some
cases such as that illustrated in Figures 4-52 through 4-55, hail contamination may affect radar
rainfall estimates. That is, when significant hail fall accompanies a storm this will tend to
artificially enhance rainfall estimates above that which actually occurs. While certain adaptable
parameters are included to mitigate the rainfall accumulation hail bias, it cannot be concluded
that all the bias is eliminated (Fulton et al. 1998).
4.12.2.1 Reflectivity. Reflectivity time lapse is a method to identify a train of cells

passing over the same area, stationary cells, or merging cells. Often, with the merger of cells,
precipitation rates associated with collision and coalescence will markedly increase.
4.12.2.2 Vertically Integrated Liquid Water. Monitoring the updated VIL

product not only provides information about water content within a storm (not precipitable water),
but a diminishing of the VIL values that might be indicative of a weakening trend. A time lapse
sequence of this product may provide this indication.
4.12.2.3 Echo Tops. Echo Tops, especially the high-resolution version, EET, can
provide information on trends in overall precipitation intensity. Increasing precipitation tops may
signal increases in updraft strength, CAPE, and/or the depth of the collision and coalescence
process. All are important to the flooding potential.
4.12.2.4 Storm Tracking Information. Movement of an intense storm or region

of storms is likely the most important factor of flash flood events. A storm with a high rainfall rate
may be moving too fast to produce a flash flood in any one location. On the other hand, a storm
with a moderate rainfall rate could cause a flash flood if it is persistent over one location or moves
along a stream or river basin. The STI product can be helpful highlighting a region of excessive
rainfall and current flooding, or in determining future areas of threat from flash flooding. If
movement is not erratic, strong individual storm cells should be adequately tracked by the product.
4.12.3 Considerations. As pointed out in the above discussion, flash flooding and flood
warnings are dependent on many factors; meteorological, radar (e.g., Z-R relationship used), and
otherwise. Thus, as is the case with all other warning types, integration of all available information
is essential for an effective flood and flash flood warning program.
Stationary cells will be easily identified with time lapses of Reflectivity and Precipitation
Accumulation products. The STI product may indicate a continuous storm cell motion but the
upstream edge of the storm complex, where cell development occurs, may be nearly stationary.
Thus, a succession of cells within the storm complex will pass over the same location, while the
complex itself is nearly stationary.
The precipitable water content of the inflow to a region of convection and heavy rain (or
potentially heavy rain) is very important. But the user must understand that VIL is a radar-

4-105
derived estimate of the liquid water content within existing storm cells or storms, but it is not an
estimate of precipitable water values.
The presence of hail in storms can enhance the reflectivity values. The Precipitation Processing
software contains logic and adaptable parameters that are designed to minimize the effects on the
estimation of precipitation amounts. However, if these parameters are not properly adjusted, hail
may cause considerable overestimation of rainfall accumulation. Other factors within the
Precipitation Processing software, such as use of the proper Z-R relationship will also affect
flood warning accuracy.
Precipitation estimates are the most critical element in making daily river forecasts to support a
multitude of hydrological decisions. These include decisions pertaining to reservoir operations for
flood control, hydropower generation, navigation, and industrial and municipal water supply
decisions. The Precipitation Accumulation products, especially the Hourly Digital Precipitation
Array (DPA), are used as input to river forecasting models. Instead of the accumulation products’
2 km (1.1 nm) x 1° polar grid, the DPA has a rectangular grid of about 4 x 4 km (2.2 x 2.2 nm).
Instead of the accumulation products’ 16 data levels, the DPA has 256 data levels. Similar to the
OHP, the DPA has a moving one-hour accumulation (scan-to-scan accumulations). The product is
used by the NWS RFCs to generate precipitation input for the NWS River Forecast System, and the
AWIPS Multisensor Precipitation Estimator (MPE) used for the bias calculation. The rectangular
grid allows for mosaicing the numerous WSR-88Ds within the RFC’s area of responsibility.
The models run by the RFCs produce numerical mosaics of precipitation accumulation that are
intended to alert the RFC forecasters to those basin areas demanding immediate attention. These
models estimate soil moisture and runoff, important to river and flood forecasting.
4.13 Non-Convective High Wind Events. The Doppler capability of the WSR-88D has proven
useful in the detection of high wind events. These events include strong gradient winds that are the
result of a deepening or a strong extratropical cyclone. Some hazards that may result from high
winds include dust storms, blowing snow, and dangerous wind-chills. Certain types of windstorms
are more terrain enhanced and are inherent to particular geographical locations, such as the Santa
Anna and the Chinook winds. Two products that can be very useful for observing and forecasting
wind speed and direction are the V and VWP.
4.13.1 Recognition of High Wind Events. The V and DV products can be used to
determine wind field characteristics when sufficient tracers are available. The speed of the winds
can be deduced by observation of the strength of the velocity field depending on tracers and the
viewing angle and, if the radar is within those winds, the VWP. Prior to the onset of a high wind
event, the velocity field may indicate stronger winds at mid-levels or lower mid-levels that could
possibly mix downward to the surface. Deep environmental lapse rates and identifying the mixed
layer are important in this regard. These higher velocities may also be evident in the VWP product,
again, depending on tracers and the wind location relative to the radar.

4-106
4.13.2 Considerations. Depending on the amount of sufficiently large scatterers or
sufficiently strong refractive index gradients in the atmosphere, the velocity field may or may not
reflect high wind events. On the average, clear air returns extend to ~30 nm in horizontal range
and ~3,000 feet in height depending on site, season, time of day, and airmass. If operating in Clear
Air Mode, and the interest is in the most accurate velocity estimates possible, selection of the short
pulse may be preferable. However, the long-pulse mode uses as many as 250 pulses per pulse
volume and is, therefore, more sensitive (~7 db more sensitive). (However, long pulse can also
induce more dealiasing failures.)
If dust storms are strong enough so that particulates are lofted into the atmosphere, a significant
echo return in the velocity field may appear. Blowing snow is often confined to a shallow depth
close to the ground and, therefore, will not likely be revealed in the velocity field except for close
to the radar.
Some lofted particulates associated with fires and smoke are rather easily detected and strong
winds are commonly associated with large fires. Therefore, monitoring fire-related windstorms
(when terrain produces only limited blockage) is a significant capability. In fact, the early
detection of fire related particulates within smoke (and/or debris) may be the first indication of
significant fire (or explosion) (Figure 4-56). The R, DR, V, DV, and the VWP products are the
primary products of choice.
4.14 Low-Level Jets. The low-level jet is the axis of a band of wind flowing at a faster rate than
the overall environmental wind within the planetary boundary layer. It may be induced by
orographic features, synoptic fronts, mesoscale, and even strong convective storm systems, or it
may result from the horizontal temperature gradient created by the diurnal heating and cooling
cycle. Low-level jets also act as conveyor belts for feeding moisture into weather systems and their
associated rain/snow or convective storms. Severe thunderstorms have been observed to develop
within, or at the nose of and downstream of, the low-level jet.
4.14.1 Recognition of Low-Level Jets. The low-level jet may be observed in velocity
fields as a couplet of strong values surrounded by weaker winds above and below, upstream and
downstream, of the radar (Figure 4-57). While not shown here, it is readily evident in the VWP
where the time history and changes in height can be observed.
4.14.2 Considerations. At further distances from the radar, the sensitivity is decreased
and the beam intercepts the atmosphere at progressively greater heights. As a result, the low-level
jet may not be depicted well, or at all, depending on scatterers in the lower elevation scans of V
products and beam height. Often, the low-level jet may be in the range of 0.9 to 1.8 km (3,000 ft to
6,000 ft) AGL. However, somewhat higher elevation scans of the velocity product may depict core
velocity regions better at close radial distances from the radar.

4-107
Figure 4-56
Smoke Plume Detected by Radar
A plume of smoke-borne particulates from a fire is detected in Clear Air Mode (top center
of graphic) by the Boston, MA WSR-88D at 03:56 UTC on 12 December 1995 (NCDC
NEXRAD Viewer graphic).

4-108
Figure 4-57
Base Velocity Data Array Product of a Low-Level Jet
Oklahoma City, OK WSR-88D Base Velocity Data Array product (DV) at 05:31 UTC on 30
May 2004 (AWIPS display). This image shows a south-southwesterly low-level jet with a
maximum velocity from 191° at about 110 kmhr-1 (60 kts).

4-109
Normally, the VWP product is the primary product of choice. One radar, assuming adequate
tracers, will depict the low-level jet over the radar. However, because of the limited width or
horizontal extent of the amplified wind field, it may not be sampled by some regional radars. If the
low-level jet is sufficiently large horizontally, then other adjacent radars may also sample the low-
level jet or “sheet” of amplified winds.
4.15 Wind Shear. A major hazard to aviation is the presence of low-level wind shear. Wind shear
may result from a variety of phenomena such as boundaries of all types, terrain, and strong
inversions. Moreover, owing to the beam width at longer ranges (e.g., ~3.5 km at ~230 km) the
normal occurrence of wind change with height will result in wind shear within the beam.
4.15.1 Recognition of Wind Shear. Wind shear along a boundary will appear as a
discontinuity in the Doppler velocity field and is frequently associated with broadened velocity
spectrum widths (Figures 4-58 and 4-59). Detection of wind shear and/or turbulence within the
pulse volume is best accomplished via spectrum widths. Therefore, assuming sufficient tracers are
present, frontal surfaces aloft or shear of the horizontal winds will be detected by enhanced areas,
or layers, in the Spectrum Width product. Spectrum Width values of 5 ms-1 (8 kts), or higher, are
usually associated with significant wind shear or turbulence, or both, within the beam.
Another method of inferring or observing wind shear is the observation of radial or

azimuthal velocity gradients. This highlights local wind-shear phenomena in larger regions, as can
the Combined Shear (CS) product. However, the CS product includes substantial smoothing which
can actually prevent detection of some of the shear. In other words, some of the shear values are
smoothed to the extent that some shear is actually eliminated. The VWP product can also be useful
in detecting and monitoring significant wind speed and direction changes (wind shear) in the
vertical within a limited radius of the radar. The default value for this radius is 29.6 km (16 nm),
but that value is adaptable. However, if that shear is too large and contributes to a large asymmetry
in calculation of the proper VAD sine wave value at any given height, then no data (“ND”) results
in the VWP product for that height.
4.15.2 Considerations. In addition to significant smoothing, the CS tends to be "noisy"

even when smoothed and, in the absence of larger and strong signatures, may be of little use. The
product does not take vertical shear into account. For indications of shear in the vertical, the VWP
or the VCS products may be useful.
4.16 Turbulence. Any disruption of ambient flow such as wind shear or flow deceleration (flow
within downdraft regions or divergence at storm summit or the earth’s surface) results in
turbulence.

4-110
Figure 4-58
Severe Weather Analysis (Velocity) Product
Norman, OK WSR-88D Severe Weather Analysis (Velocity) product (SWV) at 23:04 UTC
on 18 June 1992 (PUP display). The product shows a strong thunderstorm gust front
where bright red abruptly changes to bright green.

4-111
Figure 4-59
Severe Weather Analysis (Spectrum Width) Product
Oklahoma City, OK WSR-88D Severe Weather Analysis (spectrum width) product (SWW)
at 23:04 UTC on 18 June 1992 (PUP display). This product is at the same time as Figure
4-58. Note the broadened spectrum width values in the area of the thunderstorm gust front
(near the white cross hair).

4-112
4.16.1 Recognition of Turbulence. In the immediate vicinity of thunderstorms. for
example, when the signal-to-noise ratio is good (~10 dB), spectrum width values of 8 knots or
greater indicate turbulent areas. Higher spectrum width values will typically occur in regions of
downdraft, low-level thunderstorm outflow and gust fronts (Figures 4-58 and 4-59), and in
divergent areas at the top of storms.
Note that in Figure 4-58 the inflow to the storm (red velocities in the lower right) is rather smooth,
while in Figure 4-59 the region of the gust front (sharp change from the red velocities to the bright
green in Figure 4-58) is turbulent (note the broad spectrum widths in Figure 4-59). Note also in the
upper left of Figure 4-58 where there is a mix of inbound and outbound velocities (shear), in the
same area in Figure 4-59 denoted by broadened spectrum widths. This is a turbulent region as
might be expected.
4.16.2 Considerations. The difference between turbulent areas and strong shear regions
cannot be distinguished in the SW product because broadened SW values result from either shear
or turbulence within the beam. Additionally, regions of wind shear typically break down with
energy cascade downward into turbulence. Within a sample volume, spectrum width values may
be higher due to shear, turbulence, or noise (low signal-to-noise ratio (SNR)). Spectrum
broadening is secondarily caused by antenna rotation rate or the spread of scatterers' terminal fall
velocities. These contributions are minor.
Spectrum width values widen significantly and Doppler velocities become erratic in areas with
very low signal-to-noise ratios. In fact, the primary use of SW is now data quality. However,
because of shear of the horizontal winds with height and increasing beam width with range, SW
values broaden with increasing distance from the radar.
4.17 Microbursts. Microbursts occur when a rapidly subsiding column of air descends and
spreads out at and near the surface. Frequently, this air is negatively buoyant, but there are
exceptions when the air will overshoot its equilibrium level and continue to sink. There is no
requirement that the divergence be axial symmetric, however, and it often is not. Microbursts may
occur in the absence of precipitation when ice crystals or rain drops in virga evaporate aloft,
cooling the air rapidly, creating negative buoyancy and accelerating the air toward the surface.
Additionally, in thunderstorms, a high liquid water mass and heavy rainfall will also accelerate
downdrafts, through precipitation drag, at times becoming microbursts. There also exists a
“hybrid” class of microbursts that are created through both negative buoyancy and precipitation
drag.
4.17.1 Recognition of Microbursts. Dry microbursts are detectable within WSR-88D

velocity images due to atmospheric density gradients, insects, and particulate matter. For the “wet”
microburst, the precipitation tracers are normally abundant. The microburst is identified by a
significant velocity difference along a radial or group of radials with the radar-relative inbound
velocity at a nearer radar range than the outbound value. To qualify as a “microburst,” the distance

4-113
between the velocity maxima in the signature must be less than 4 km apart. When the distance
between these maxima are larger the signature qualifies as a “macroburst.”
Thus, a microburst will produce a unique signature detectable at low levels in the velocity field. A
divergence couplet (i.e., a pattern of “strong” flow toward the radar, matched by an opposite
pattern of “strong” flow away from the radar) will be seen on the display (Figure 4-60). A
microburst is located in Figure 4-60 at the white cross where the velocity divergence is 23.7 ms-1
(46 kts), 10.3 ms-1 (20 kts) inbound, and 13.4 ms-1 (26 kts) outbound. Divergence of this
magnitude is frequently damaging (Burgess and Lemon 1990). The strength of the signature is
highly dependent on radar range to the microburst, as well as the angle of observations. The
divergence signatures are often asymmetric. At times, apparent asymmetry can be created by
storm motion superimposed on the microburst’s outflow. In this situation, either the velocity
toward or away from the radar will be higher than its opposite. The strong downflow from a
microburst descends very close to the ground before diverging. This divergence in the velocity
field might also be detectable in VCS products if the signature is close enough to the radar (Figures
4-61 and 4-62). When descending air is also translating along the radial, the resulting divergence
signature will most likely be asymmetric. But in many situations microburst signatures will not be
symmetric in any case. Divergence maximizes from 2 to 4 minutes after initial development of the
signature with a diameter of approximately 4 km (2.5 nm) or less and a velocity difference
sometimes exceeding 51 ms-1 (100 kts). However, while divergence signatures of this strength are
very rare, Burgess and Lemon (1990) found that a WSR-88D signature with a velocity difference
of 18 ms-1 (35 kts) or more are generally damaging. The most intense velocity difference in the
divergence signature is normally below about 300 m (~ 980 ft). For this reason, these local and
sometimes damaging “outbursts” of wind are often below the radar horizon and, therefore, invisible
to the radar unless the microburst is within ~ 40 km.
The products of choice for detecting and monitoring these signatures are R, DR, V, DV, and SWV,
and the RCS and VCS. Note that these velocity products are ground-relative instead of storm-
relative. Obviously, the damaging flow is ground-relative, thus these are appropriate products.
4.17.2 Considerations. As stated above, with the shallow vertical extent of the outflow
from a microburst, this phenomenon will usually not be detected beyond the range of about 20 nm.
In addition, because of the short-lived nature of the microburst phenomenon (typically 3 to 5
minutes) and the current scanning strategies used in the WSR-88D, Microbursts may be
infrequently detected at their maximum divergent intensity. However, VCP 12 is the coverage
pattern of choice because of the overlapping low-level beams in the vertical and the relatively rapid
update rate (4.1 minutes).

4-114
Figure 4-60
Microburst Detected on a Mean Radial Velocity Product
Oklahoma City, OK WSR-88D Mean Radial Velocity product at 17:16 UTC on 5 June 1991
(PUP display). A microburst is located at the white cross hair where the velocity
divergence is 23.7 ms-1 (46 kts), 10.3 ms-1 (20 kts) inbound and 13.4 ms-1 (26 kts)
outbound.

4-115
Figure 4-61
Reflectivity Cross Section of a Microburst-Producing Thunderstorm
Phoenix, AZ WSR-88D Reflectivity Cross Section product at 00:29 UTC on 15 August 1995
(AWIPS display). This product shows a microburst-producing thunderstorm shortly after the
feature reached the surface.

4-116
Figure 4-62
Velocity Cross Section of a Microburst-Producing Thunderstorm
As in Figure 4-61 except for a Velocity Cross Section product. Note the convergence aloft
(~7,000 ft) into the reflectivity core and the divergence signature below extending just
above the radar horizon.

4-117
At ranges nearer the radar (~ 40 km and less), the DV may prove most useful. At further
distances from the radar, microbursts may be anticipated by monitoring high reflectivity cores at
mid-levels 5 to 12 km (16,000 to 40,000 feet) as they descend and are often accompanied by
flow convergence into these cores. Additionally, observations of the DCZ and the MARC
regions are also indicative of both macrobursts and microbursts, but often the lead time for the
MARC is generally on the order of two volume scans (Schomocker et al. 1996). Additionally,
the false alarm rate may be high. Collapsing echo tops or rapidly diminishing VIL values may
also indicate such an event. However, the collapse of some storms results in only weak outflow.
Additionally, gaps in the VCPs may themselves create the apparent “collapse” (Howard et al.
1997; Maddox et al. 1999).
4.18 Considerations for Building a Routine Product Set. The following information is intended
as guidance to users setting up RPS Lists and is subject to modification by agency-unique
operational requirements. The data display systems typically provide the user with the capability to
define several unique RPS Lists.
4.18.1 Initial Considerations. Each Associated User typically has the capability of
creating and accessing numerous RPS lists depending on the type of weather situation. Every
RPS List must be designed carefully in order to minimize the amount of editing of the “Current
List.” Included here are some of the considerations for RPS List determination. These are
guidelines only. The considerations below apply typically for 33.6 – 56 Kbps line speeds.
4.18.2 Some Local RPS List Considerations. Routine Product Set Lists must be
designed with several things in mind. The object of the list is to provide the user with products
needed most often in order to best sample the atmosphere and thereby interpret the current weather
situation. Things which must be taken into consideration are:
1. Weather Mode/VCP Clear Air? Precipitation? Which VCP? You need to anticipate
this information in order to know from what angles you have to
choose.
2. Weather Situation Stratiform? How widespread will the precipitation be?

Convective? Low Topped? Winter Weather? Clear Air Return?
The user needs to know this to determine which products and what
elevation angles will be utilized.
3. Range Where are the echoes located? Close to the RDA? Beyond 124 n
mi? Both? This information will determine which of the current
elevation angles will sample the precipitation at the appropriate
height, considering the range to the echo and the anticipated
vertical extent of the echoes being sampled.

4-118
4. Procedures Finally, the RPS List needs to be compatible with your display
procedures. For example, when planning use of multi-panel
displays, the proper products and product elevation angles must be
on that list. Often, however, it will be the procedures which will
be designed around the RPS Lists.
Associated Users with a 14.4 Kbps and 33.6 Kbps lines are limited to 50 products. Line quality
may reduce the number of products available. Users with slower lines must also anticipate that the
larger products such as SW or CR products also reduce the number of products received each
volume scan.
4.18.3 Product Considerations. No predetermined RPS List, outside the National List,
will always be appropriate, no matter how carefully thought out it is. Some overlap is necessary in
order to keep real-time editing at a minimum. The user must always be prepared to change and
fine tune the RPS List to ensure the products received are the products needed. Also, the user must
keep in mind how these changes will affect the procedures used.
It is not always easy to decide which products to choose. Here are some considerations:
1. At least 4 elevations of reflectivity and velocity (Storm Relative) are needed at a

minimum for proper storm structure evaluation. The elevations chosen will depend
on the VCP, range, and the vertical extent of the anticipated echoes. For storm
interrogation, products representing low, middle, and storm summit locations
should be viewed as a minimum. As line speed permits, even more elevation cuts
are desirable.
2. Additionally, the higher resolution base products such as DR and DV should

always be considered, especially if severe weather is anticipated. Thus, the user
should consider placing several slices of the high-resolution Velocity on the RPS
List, as well as the matching DR products. This will be especially useful if
features are small in scale or at a significant distance from the radar (i.e., being
severely impacted by aspect ratio). It will also be beneficial with boundary
detection or if echoes are of weak intensity. Also, the user will need this product if
it is required to view the high resolution SRM display. The user must be aware
that, for other than LAN-to-LAN connections, load shedding could be an issue.
The same can be said for the high-resolution Reflectivity products.
3. Velocity products should never be obtained without a matching Reflectivity (or

high resolution Reflectivity) product. Both types of base data are needed for
proper interpretation.
4. The SRM and SRR subtracts the average motion of all detected storms (or storm
for the region product) from the Base Velocity product so that flow relative to the

4-119
storms is presented. These products are best utilized when circulations are
expected and storms are fast moving. The user must keep in mind that the SRM is
inappropriate if ground-relative winds are of interest, like those associated with
gust fronts and high wind events. The DV or DV products are more suited for
ground-relative wind situations.
5. Some display systems may construct base velocity/and base reflectivity displays
using the best resolution data that are available in the database. This is desirable.
Therefore, if higher resolutions are present in these displays, they must be included
on the RPS List. For example, if only a 4 km (2.2 nm) resolution CR product is
required in the database, the product displayed will have only that resolution. If,
however, both a 1 km (0.54 nm) and a 4 km (2.2 nm) resolution product are in the
database, then the CR display will show a combination of both resolutions, using 1
km (0.54 nm) out to 230 km (124 nm), and 4 km (2.2 nm) from 230 to 460 km
(124 to 248 nm).
6. MESOCYCLONE and TVS (M/MD, TVS) products should be included on all

RPS Lists anytime convection is expected. They typically need to be checked
every volume scan or displayed with other products (e.g., SRM) to prevent any
unanticipated detections from being missed. When detections are indicated the
user might consider adding the rapid update product versions (MRU and Tornado
Vortex Signature Rapid Update (TRU)). However, note that TRU is selected on an
elevation-by-elevation basis, so it can push the RPS List to its limitation on the
number of entries. Note: The TVS and M/MD products should also be set for
alerts for the same reason, keeping in mind that for RPG alerts, only the first (of
many continuous) detections will be alerted until there is a volume scan free of
either detection.
7. The STI product is useful out to 186 nm and may, therefore, be valuable on a Clear
Air RPS List to monitor storms outside of 124 nm.
8. Precipitation algorithms will run while the radar is still in Clear Air Mode under
certain conditions. Therefore, the user may choose to include precipitation
products on the Clear Air RPS List.
4.18.4 National RPS List Requirements. To support the central collection and
distribution of radar products via the RPCCDS, NOAAPORT, and the AWIPS Satellite Broadcast
Network (SBN) Wide Area Network (WAN), a national set of products (list of products is at:
http://www.nws.noaa.gov/tg/rpccds.html) is collected from 155 WSR-88D systems. For these
sites, the National Product Set also includes Archive Level III products (Part A, Chapter 5, of this
Handbook) and others identified for national collection.

4-120
Whenever a new RPS List is sent to the RPG, either automatically through a VCP mode change or
manually through a request sent from the display system workstation (via the RPS List), the
National Product Set is combined with the local user-defined RPS List and resent to the RPG.
Even if the user needs to modify the “Current” RPS List or build a new list, it will be combined
with the National Default List for the current operational mode. Duplicate products on the
combined list will be dropped if a National product has a similar elevation angle (within 0.2
degrees) of the local office list and all other parameters are the same.
The site controlling the MSCF is responsible for modifying and copying the local user-specified
RPS Lists to accommodate the changes brought about by the national central collection lists
together with local needs. The differences in the various lists are, in general, due to the differences
in clear air versus precipitation VCPs and those brought about by varying line speeds.
4.18.5 Clear Air Mode. Due to the slow scan rate, most of the products on the RPS List
can be base products. The user may include a 16-level, low-elevation Clear Air Reflectivity
product and an 8-level, long range Reflectivity product. Such derived products as the VWP and ET
may also be added. Generally, there is time to handle OTRs for other products. Note that, in
almost all cases, the VWP need not be requested every volume scan. In fact, in clear air conditions
the rate of changes in the atmosphere may be slow enough that many of the products have request
rates less often than every volume scan.
4.18.6 Precipitation Mode. Building RPS Lists for VCPs 11, 12, 121, and 21 can be
complex. As explained above, conditions that influence the selection of products are the type,
severity, range, and aerial extent of the precipitation. In addition, the support being provided to
other offices, users, and WSR-88D sites may be a factor in building the RPS List.
With severe storms occurring at long range the RPS list should concentrate on low-elevation base
products at the highest resolution. The STI product may be added. If storms are between 62 nm
and 124 nm, velocity-derived products such as M, MRU, MD, and SRM may be included. Since
the data compaction techniques are more efficient when the area of coverage is small, the RPS list
can consist mainly of base products without inducing load shedding. One-time requests for such
products as Cross Sections, SWA, and SRR may also be accomplished without inducing load
shedding. (The use of VCP 121 and 21 will allow the user more time to make OTRs, as well as
providing good resolution of base products.) The lowest five elevation scans for VCPs 11, 121,
and 21 are identical in supporting the hydrology algorithms and will cover storms at long range.
However, consideration should be given to VCP 12 use for longer ranges because of low-level
overlapping beams.
An RPS List for convective precipitation at short to medium ranges should be built around VCPs
11 and 12. This will optimize algorithm performance and provide intermediate elevation scans for
detailed analyses of storms using the SWA products and multi-panel base products. The RPS List
should include one long range Reflectivity or Composite Reflectivity product and a few base
products at intermediate elevation scans. About half the RPS List may consist of algorithm-derived

4-121
products. However, sufficient base products must be included to allow manual identification of
severe storm circulations and structure and to allow the user to check algorithm output against base
data. Again, products such as VWP need not be requested every volume scan.
If precipitation is widespread, base product size will be large and require longer transmission times.
In this case, and if severe convection is not anticipated, the RPS List should be composed of fewer
than one-half base products. The Precipitation Accumulation, ET, VWP, CR, and low-elevation
angle R and V products may be included. Other base products may be needed to monitor middle
and upper level atmospheric changes. But, in all cases, the user should rely on base data analysis.
4.19 Automated Alerts. With the WSR-88D, the user will be receiving large amounts of data for
each volume scan. Since it is impossible to monitor all of these products at all times, it is
sometimes helpful to be notified when certain predetermined meteorological criteria are met or
exceeded within data from each volume scan. It is in this way that the alert process, which can be
specified at many workstations, can help the user monitor this large volume of data and avoid
overlooking critical data (Tables 4-1 and 4-2). The user can choose to set alerts using the
appropriate equipment and procedures. Here we discuss, as an example only, one way this might
be done within agency offices. Specific agencies or offices may choose to use the alerts in another
fashion, or these agencies may have overriding concerns and directives.
As an example, widespread showers cover the screen but nothing severe is indicated. The
user/forecaster will likely be devoting most of his/her time to other duties. If the user sets the
reflectivity threshold at 50 dBZ, he/she will receive an alert when that criterion is first met or
exceeded within the specified alert area or forecast to move into the area. This type of usage will
supplement the users’ monitoring of marginal weather. Or, while viewing a magnified display
centered on an intense storm in one portion of the area, a mesocyclone is detected in another storm
not currently on the display. The alert feature in this instance could notify the user of potentially
severe storms not currently in view.
Once the user establishes what thresholds are critical for their area and mission for a given time of
year, or for a particular weather scenario, he/she will be able to set alerts for these thresholds and
determine in what part of the radar coverage area these thresholds will be monitored (office area of
warning responsibility, for example).
4.19.1 Alert Use. Alerts in the WSR-88D are best utilized in two different ways: 1)
During the transition from a "non-warning" or “monitor” type mode, to a "warning" situation, or 2)
While in a warning situation, to highlight very significant weather that might be otherwise
overlooked due to the high workload, or simply because this threat type was not anticipated. Thus,
alerts can increase situation awareness. Alert use for marginal storms is not as valuable once
widespread or frequent severe weather has begun. Below are some factors in the alert process
design which affect their use in other instances.

4-122
1. The alert process will notify the user of an "alert condition.” This means that once
a phenomenon (in a particular alert area) has been alerted for, the user will not
receive any additional alerts for that phenomenon until a volume scan passes in
which that particular criteria is not met by any storm in the alert area. In other
words, if one storm sets off the alert, the user will not get a new alert for that
criterion when any other storm reaches the threshold, (even if the later storm is
more severe) because the "alert condition" has not changed. In the case of TVS
detection, for example, this means the user can not rely on the alert process to
notify him/her of each TVS detection in the first volume scan where criteria are
met, nor can the user expect any alerts for TVSs on subsequent volume scans if
additional ones are detected. The user must look at each TVS/TRU product (and
it’s assumed the user would be) in order to observe detections at this point.
2. In general, if several storms in the same alert area contain algorithm-identified

mesocyclones (for example), only the first one detected will be listed on the alert
status screen.
3. The user will not get alerts for most features until after the entire volume scan is
completed, since most features require the entire volume scan in order to be
identified. This may be several minutes after the volume scan start time. The user
may, however, get the alert itself before the product (which may display the alerted
feature) is received into the user’s database, since alerts have a higher priority than
products. For example, a MESO alert could be sent to his/her workstation before
the MESO product containing that feature arrives.
The net result of these characteristics makes relying on the alert process to notify the user of all
severe storms a dangerous practice. Alerts are not designed for that purpose. Algorithm output
and associated alerts should not replace the practice of monitoring base data and using this data for
storm investigation/interrogation. The receipt of an alert should only serve to spur the user into
further investigation and validation of the alerting feature. Furthermore, under no circumstances
should the absence of an alert or algorithm-identified signature imply the absence of either severe
weather or the existence of said signature. For this and other reasons, this document emphasizes
reliance on base products.
4.19.2 Setting Alerts. Even though there are serious limitations to the alert process, an
effective strategy can still make good use of this process. For example, VIL will trigger with a
value of 50, TVS will trigger with the detection of a TVS, and MESO will trigger with the
detection of a mesocyclone. One-Hour Precipitation and CR are also set. These alerts settings are
being used in a “monitor” mode. But this does not alleviate the user of their responsibility to
manually monitor and examine base data. Should this user get detection and shift gears into more
of a warning mode, then the user might raise the threshold on VIL and CR, for example, to alert for
the high-end detections as this user would no longer need notice of marginal thresholds. For each
of these, “no” is the setting for alert-paired products in the example. Values assigned to the

4-123
threshold codes are fixed at the RPG and can only be altered with URC permission. The user’s
option is merely to select which of the established codes to choose for his/her use.
This strategy allows the user to be notified when marginal echoes develop, and then modify
thresholds to get alerts for the most serious, and sometimes overlooked, features once warning
operations begin. However, the user should also weigh the real need for alert use when the user is
now applying their skill to storm interrogation through base data examination. The user must also
keep in mind that if many alerts are set, the messages and/or audible alarms will probably become a
nuisance, and will likely be ignored or turned off, thus defeating the purpose of the alerts and the
alarm. On the other hand, the user does not want significant occurrences to go by undetected
(excessive precipitation accumulation and flood potential during tornado occurrence, for example).
The user must remember the limitations of the Alert Process: there will not be new alerts for
categories where "alert conditions" are still in effect until one volume scan has passed without that
condition being detected anywhere in the alert area.
4.19.3 Alert Groups. Alert categories are divided into 3 alert groups, defined as follows:
GRID GROUP - Phenomena within this group meets or exceeds the thresholds in a
specific Cartesian grid box. For example, a VIL value of 50 can be traced to a particular
grid box.
VOLUME GROUP - Phenomena within this group meets or exceeds the thresholds for
algorithm output assigned to a particular storm. For example, the MEHS refers to the
storm as a whole, not necessarily to a specific grid box affected by the storm.
FORECAST GROUP - Phenomena within this group are currently associated with storms
forecast to move into the established Alert Area. When the alert threshold is met and the
"associated" storm is forecast (by the SCIT algorithm) to move into an Alert Area, a
"forecast" alert is generated. Note - the forecast alert will be generated when a feature
outside the alert area meets criteria and is forecast to move into the alert area. However, a
forecast alert will also be generated if a feature within the alert area is forecast to remain
within the alert area for at least one forecast period (15 minutes). This may thereby limit
the usefulness of the forecast alerts once storms move within the alert area (assuming that
the same criteria are already being alerted for within the alert area via the Grid or Volume
alerts).
Additionally, the user must recognize that the alert algorithm does not actually anticipate
development of a condition that does not already exist within the storm. The only thing actually
“forecast” is storm movement, and that is done by the SCIT algorithm.
4.19.4 Alert Areas and Acknowledgments. Alert Areas are often set up to represent
warning areas. One may overlap the other, be totally surrounded by another, or be totally

4-124
contained within another. In addition, a single alert area can be comprised of one area the size of
the entire screen, or of several smaller areas which are non-contiguous. When alerts are received at
a workstation, these alerts must be acknowledged. Otherwise, either/or audible and visual alarms
will continue.
4.19.5 User Alert Message (UAM) Product. The UAM does not need to be on the RPS
List in order to be received. Whenever alerts are set, the RPG will automatically produce this
product every volume scan to indicate which new alerts were triggered during the current volume
scan. The UAM is displayed on a text workstation along with the 3-letter radar mnemonic. Since
this message contains only new alerts for the current volume scan, it is possible to have outstanding
alerts in effect while the content of this product displays “No New Alerts This Scan.”
4.19.6 Alert-Paired Products. Some alerts may also result in products being sent to the
user workstation. The user may choose to receive an RPG-specified product at the same time the
alert is generated. These products are pre-assigned at the RPG and are also approved at the URC
level of change authority. The user may only choose whether or not to receive these products. The
product chosen to arrive with a particular alert may be specified so as to provide additional
information about the phenomenon. The user may display the product from the menu just as any
other products are displayed. It is, therefore, useful to know what the paired products are for
various alerts in order to recognize them when they arrive. For example, a MESO Alert generated
when a mesocyclone is detected may come with an SRR centered on the area where the
mesocyclone was detected. Some example pairings for alerts which have proved effective are: For
both the MESO and the TVS, the Storm Relative Region product, which will be generated at the
slice and range of the feature triggering the alert. This reasoning is assuming that BOTH TVS and
M are already on the RPS List. There is no advantage to receiving the “paired product” if that
paired product is on your RPS List.
Alert-Paired Products are given the highest priority and will take up space in the queue and time on
the transmission lines. If the user is having recurring problems with narrowband load shedding the
way to help alleviate this problem is to decrease alerts or decrease or eliminate paired products at
your workstation. However, alert-paired products are sent so rarely that this is insufficient to
produce this problem. Other products may need to be deleted instead.

4-125
Table 4-1
Alert Types and Mnemonics (With the Assigned Order of Importance)
Alert Alert Group Alert Type Mnemonic Order of

Category Importance
1 Grid Velocity GV 19
2 Grid Composite Ref. GR 13
3 Grid Echo Tops GT 16
4 Grid Severe Wx Prob GP 21
5 Grid Not Used
6 Grid VIL GL 7
7 Volume VAD VD 20
8 Volume Max Hail Size VZ 8
9 Volume Mesocyclone VM 3
10 Volume TVS VS 1
11 Volume Max Storm Ref VR 14
12 Volume Prob of Hail VH 10
13 Volume Prob Svr Hail VA 5
14 Volume Storm Top VT 17
15 Volume Max 1-hr Precip VP 12
16-24 Volume Not Used
25 Forecast Max Hail Size FZ 9

26 Forecast Mesocyclone FM 4
27 Forecast TVS FS 2
28 Forecast Max Storm Ref FR 15
29 Forecast Prob of Hail FH 11

4-126
Table 4-2
Alert Locations and Paired Window Product Centers
CATEGORY CODE ALERT AZ/RAN ,SLICE LOCATION

(as seen on S, AL screen)
GRID GROUP
1. Velocity Lowest Elevation &/or AZ/Ran of alert box
containing exceeding value.
2. Composite Reflectivity " "
3. Echo Tops " "
4. SWP " "
5. -
6. VIL " "
VOLUME/FORECAST GROUPS
7. VAD / VWP Elevation computed from height and slant
range of VAD AZ/Ran.
15. MAX 1-hr Rainfall RDA
8, 25 HAIL Storm Centroid AZ/Ran
9 MESO (Volume) Meso: Feature Base elevation and AZ/Ran
2-D, 3-D Shear: Storm Centroid AZ/Ran and
Lowest Elevation
25 MESO (Forecast) MESO, 2-D, 3-D: Storm Centroid AZ/Ran
and Lowest Elevation
10, 27 TVS AZ/Ran and elevation of base
11, 28 MAX Storm Reflectivity Elevation of MAX, Storm Centroid AZ/Ran
12, 29 MAX Storm Velocity Elevation and AZ/Ran of lowest storm
component.
13, 30 Storm Volume Storm centroid AZ/Ran, and Lowest Elevation
14, 31 Storm Top Storm centroid AZ/Ran, and Lowest Elevation
PAIRED PRODUCT ALERT PRODUCT CENTERS (and special conditions)

SWA,SRR Grid/Volume (Meso, TVS only): AZ/Ran, elevation of highest shear
Grid/Volume (all others), and Forecast: AZ/Ran of Centroid,
Lowest elevation.
RCS,VCS AZ/Ran of Grid box, +/- 50km (generated along radar Azimuth)

4-127
Table 4-3
Application versus Product
Application Product
Boundaries R, SW, V, preferably at high resolution
Cloud/precip. Layers R. RCS, V, VWP
Convection CR, DSP, ET, HI, LRM, M, MRU, MD, OHP, R. RCS, SRM, SRR, SS,
STI, STP, SWA, SW, THP, USP, TVS, TRU, V, VIL, Time-lapse of base
R, High resolution for all products
Extratropical Cyclones CR, DSP, ET, HI, OHP, R. RCS, STI, STP, SW, THP, USP, V, VCS,
VWP, Time-lapse of base products.
Flash Floods DSP, ET, LRM, OHP, R. STI, THP, STP, USP, VIL, VWP, time lapse of
base R and OHP and STP products.
Hail DSP, HI, OHP, R. RCS, SS, STI, STP, SRM, SRR, THP, V, VCS, VIL,
High resolution products, Time-lapse of base R.
High Wind Events V, VWP, R
Low-Level Jets V, VWP
Melting Level R. RCS, SW, VWP
Microbursts CS, ET, SW, V, VCS, VIL, SRM, all high resolution products
Pre-Convective R, SW, V, High resolution products, Time-lapse loops.

Development
Tropical Cyclones ET, M, OHP, R. SRM, STI, SW, THP,V,VCS,VIL,VWP, USP
Turbulence V, SW, VCS
Wind Shear CS, SW, V, VCS, VWP
Winter Storms DSP, R, RCS, OSD, SSD, USD, STI, STP, SW, OSW, SSW, USW, V,
VCS, Time-lapse of R and snowfall products

4-128
REFERENCES
Amburn, S. and P. Wolf, 1997: VIL density as a hail indicator. Wea. Forecasting, 12, 473-478.
Atlas, D. (Ed), 1990: Radar in Meteorology. Amer. Meteor. Soc., Boston, MA, 323 pp.
Battan, L. J., 1959: Radar Meteorology. University of Chicago Press, 161 pp.
Blaes, J. L., C. S. Cerniglia Jr, and M. A. Caropolo, 1998: VIL density as an indicator of hail
across eastern New York and western New England. Eastern Region Technical Attachment,
No 98-8, National Weather Service, NOAA, Department of Commerce, Bohemia, NY
Brown, R. A., D. W. Burgess, and K. C. Crawford, 1973: Twin tornado cyclones within a severe
thunderstorm: Single Doppler radar observations. Weatherwise, 26, 63-69.
Brown, R. A., L. R. Lemon, and D. W. Burgess, 1978: Tornado detection by pulsed Doppler radar.
Mon. Wea. Rev., 106, 29-38.
Browning, K. A., 1977: The structure and mechanism of hailstorms. Hail: A Review of Hail
Science and Hail Suppression, Meteor. Monogr., 16, 1-43.
Brooks, H. E., and J. P. Craven, 2002: Database proximity soundings for significant severe
thunderstorms, 1957 – 1993. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX,
Amer. Meteo. Society, 639-642.
Bruderer, B., 1977a: The study of bird migration by radar. Part I: The technical basis.
Naturwissen-schaften, 84, 1-8.
Bruderer, B., 1977b: The study of bird migration by radar. Part II: Major achievements.
Naturwissen-schaften, 84, 45-54.
Bruderer, B., 2000: Three decades of tracking radar studies on bird migration in Europe and the
Middle East. Migrating Birds Know No Boundaries, International Seminar (and book) on Birds
and Flight Safety in the Middle East, Israel, April 25-29, 1999, Ed. Yossi Leshem, Yale
Mandelik, Judy Shamoun-Barnes. International Center for the Study of Bird Migration, Latrun,
Israel, 107-141.
Burgess, D. W., and L. R. Lemon, 1990: Severe thunderstorm detection by radar. Radar in
Meteorology (D. Atlas, Ed.), Amer. Meteor. Soc., 619-647.
Burgess, D. W., and M. A. Magsig, 1998: Recent observations of tornado development at

near range to WSR-88D radars. Preprints, 19th Conf. on Severe Local Storms,
Minneapolis, MN, Amer. Meteor. Soc., 756-759.

4-129
Burgess, D. W., M. A. Magsig, J. Wurman, D. C. Dowell and Y. Richardson, 2002: Radar
observations of the 3 May 1999 Oklahoma City tornado. Wea. Forecasting, 17, 456-471.
Burgess, D. W., R. R. Lee, S. S. Parker, D. L. Floyd, and D. L. Andra, Jr, 1995: A study of mini-
supercells observed by WSR-88D radars. Preprints, 27th Conf. on Radar Meteorology, Vail,
CO, Amer. Meteor. Soc., 4-6.
Burgess, D. W., V. T. Wood, and R.A. Brown, 1982: Mesocyclone evolution statistics. Preprints,
12th Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 422-424.
Byers, H. R., and R. R. Braham, 1949: The Thunderstorm. U.S. Government Printing Office,
Washington D.C., 287 pp.
Carven, J. P., H. E. Brooks, and J. A. Hart, 2002: Baseline climatology of soundings derived
parameters associated with deep, moist convection. Preprints, 21st Conf. on Severe Local
Storms, Amer. Meteo. Soc., San Antonio, TX, 643-646.
Cerniglia, C. S., and W. R. Snyder, 2002: Development of warning criteria for severe pulse
thunderstorm in the northeast United States using the WSR-88D. Eastern Region Technical
Attachment, No 2002-03, National Weather Service, NOAA, Department of Commerce, 14
pp., Bohemia, NY.
Chisholm, A. J. and J. H. Renick, 1972: The kinematics of multicell and supercell Alberta
hailstorms. Alberta Hail Studies, 1972, Research Council of Alberta Hail Studies Rep.No. 72-
2, 24-31.
COMET, Mesoscale Convective Systems: Squall lines and Bow Echoes, 1999, Cooperative
Program for Operational Meteorology, Education and Training, Distance Learning Program.
http://www.meted.ucar.edu/convectn/mcs/
Doswell, C. A. III, 1985: The operational meteorology of convective weather, Vol. II: Storm scale
analysis. NOAA Tech. Memo. ERL ESG-15, 240 pp.
Doswell, C. A., III, and D. W. Burgess, 1993: Tornadoes and tornadic storms: A review of
conceptual models. The Tornado: Its Structure, Dynamics, Prediction, and Hazards. (C. Church
et al (Eds), Geophys. Monogr., Amer. Geophys. Union, 161-172.
Doswell, C. A. III, Ed., 2001: Severe convective storms - An overview. Meteor. Monogr., 50, Ed.
C. A. Doswell, Amer. Meteor. Soc., Boston, MA, pp. 1-26.
Doviak, R. J. and D. S. Zrnic, 1984: Doppler Radar and Weather Observations. Academic Press,
Inc., Orlando, FL, 458 pp.
Edwards, R., and R. L. Thompson, 1998: Nationwide comparisons of hail size with WSR-88D
vertically integrated liquid water and derived thermodynamic sounding data. Wea. Forecasting,
13, 277-285.

4-130
Fulton, R., J. Breidenbach, D. J. Seo, D. Miller and T. O’Bannon, 1998: The WSR-88D Rainfall
Algorithm. Wea. Forecasting, 13, 377-395
Grant, B. N., R. Prentice, 1996: Mesocyclone characteristics of mini-supercell thunderstorms.

Preprints, 15th Conf. on Weather and Forecasting, Norfolk, VA Amer. Meteor. Soc., 362-365.
Green, D. G. and R. A. Clark, 1972: Vertically integrated liquid water. A new analysis tool. Mon.
Wea. Rev., 100: 548-797.
Howard, K.W., J.J. Gourley, and R.A. Maddox, 1997: Uncertainties in WSR-88D measurements
and their impacts on monitoring life cycles. Wea. Forecasting, 12, 166-174.
Kelly, T. A., R. Merritt, T. J. Donalds, and R. L. White, 2000: The avian hazard advisory system.
Migrating Birds Know No Boundaries, International Seminar (and book) on Birds and Flight
Safety in the Middle East, Israel, April 25-29, 1999, Ed. Yossi Leshem, Yale Mandelik, Judy
Shamoun-Barnes. International Center for the Study of Bird Migration, Latrun, Israel, 101-105.
Lee, B. D., and R. B Wilhelmson, 1997: The numerical simulation of non-supercell tornadogenesis.
Part I: Initiation and evolution of pretornadic mesocyclone circulations along a dry outflow
boundary. J. Atmos. Sci., 54, 32-60.
Lemon, L. R., R. J. Donaldson, Jr., D. W. Burgess, and R. A. Brown, 1977: Doppler radar application
to severe thunderstorm study and potential real-time warning. Bull. Amer. Meteor. Soc., 58, 1187-
1193.
Lemon, L. R., 1978: On the use of storm structure for hail identification. Preprints, 18th Conf. on
Radar Meteorology, Boston, MA, Amer. Meteor. Soc., 203-206.
Lemon, L. R., and C. A. Doswell III, 1979: Severe thunderstorm evolution and mesocyclone structure
as related to tornadogenesis. Mon. Wea. Rev., 107, 1184-1197.
Lemon, L. R., 1980: Severe thunderstorm radar identification techniques and warning criteria. NOAA
Tech. Memo. NWS NSSFC-3, 60 pp.
Lemon, L. R., and E. M. Quoetone, 1994: Interpretation of the radar-centered "Donut" Signature.
Postprints, First WSR-88D User's Conf., NOAA, Gov. Printing Office, Norman, OK, 102-111.
Lemon, L. R., and S. Parker, 1996: The Lahoma storm deep convergence zone: Its characteristics and
role in storm dynamics and severity. Preprints, 18th Conf. on Severe Local Storms, Boston, MA,
Amer. Meteor. Soc., 70-75.
Lemon, L. R., 1998: The radar “three-body scatter spike”: An operational large hail signature. Wea.
Forecasting, 13, 327-340.

4-131
McCaul, Jr., E. W., 1987: Observations of the Hurricane "Danny" tornado outbreak of 16 August
1985. Mon. Wea. Rev., 115, 1206-1223.
Maddox, R. A., Zaras, D. S., MacKeen, P. L., Gourley, J. J., Rabin, R., Howard, K. W., 1999: Echo
height measurements with the WSR-88D: Use of Data from One Versus Two Radars, Wea.
Forecasting, 14, 455-460.
Miller, D. J., and R. H. Johns, 2000: A detailed look at extreme wind damage in derecho events.
Preprints, 20th Conf. on Severe Local Storms, Orlando, FL, Amer. Meteor. Soc.,
52-55.
Miller, D. J., Andra, D. L., Evans, J. S., and R. H. Johns, 2002: Observations of the 27 May 2001
high-end derecho event in Oklahoma. Preprints, 21st Conf. on Severe Local Storms, San
Antonio, TX, Amer. Meteor. Soc., 13-16.
Moller, A. R., C. A. Doswell, and R. W. Przybylinski, 1990: High precipitation supercells: a

conceptual model and documentation. Preprints, 16th Conf. on Severe Local Storms,
Kananaskis Park, Alberta, AB, Amer. Meteor. Soc., 52-57.
Moller, A. R., 2001: Severe local storms forecasting. Chapter 11, Severe Convective Storms,
Meteor. Monogr 50, Ed, C. A. Doswell, III, Amer. Meteor. Soc., Boston, MA. pp 433-480.
Moore, J. T. and M. A. Kaster, 1993: Physical processes organizing wintertime frozen precipitation
events in the Midwest. Preprints, 13th Conf. on Weather Analysis and Forecasting, Vienna,
VA, AMS, 512-515.
Petersen, W. A., L. D. Carey, S. A. Rutledge, J. C. Knievel, N. J. Doesken, R. H. Johnson, T. B.

McKee, T. Von der Haar, J. F. Weaver, 1999: Mesoscale and radar observations of the Fort
Collins flash flood of 28 July 1997. Bull. Amer. Meteor. Soc., 80, 191-216.
Przybylinski, R. W., and W. J. Gery, 1983: The reliability of the bow echo as an important severe
weather signature. Preprints, 13th Conf. on Severe Local Storms, Tulsa, OK, Amer. Meteor.
Soc., 270-273.
Przybylinski, R. W., 1995: The bow echo: Observations, numerical simulations, and severe
weather detection methods. Wea. Forecasting, 10, 203–218.
Przybylinski, R. W., G. K. Schmocker, and Y. J. Lin, 2000: A study of storm and vortex
morphology during the ‘intensifying stage’ of severe wind mesoscale convective systems.
Preprints, 20th Conf. on Severe Local Storms, Orlando, FL, Amer. Meteor. Soc., 173-176.
Quoetone, E. M., K. Huckabee, 1995: Anatomy of an effective warning: Event anticipation, data
integration, feature recognition. Preprints, 14th Conf. On Weather Analysis and Forecasting,
Amer. Meteor. Soc., Dallas, TX, 420-425.

4-132
Shields, M.T., R.M. Rauber, and M. K. Ramamurthy, 1991: Dynamical forcing and mesoscale
organization of precipitation bands in a Midwest winter cyclonic storm. Mon. Wea. Rev., 119,
936- 964.
Schmocker, G. K., R. W. Przybylinski, and Y. J. Lin, 1996: Forecasting the initial onset of
damaging downburst winds associated with a Mesoscale Convective System (MCS) using the
Mid-Altitude Radial Convergence (MARC) signature. Preprints, 15th Conf. on Weather
Analysis and Forecasting, Norfolk, VA, Amer. Meteor. Soc., 306–311.
Spratt, S. M., D. W. Sharp, P. Welsh, A. C. Sandrik, F. Alsheimer and C. Paxton, 1997: A WSR-
88D assessment of tropical cyclone outer rainband tornadoes. Wea. Forecasting, 12, 479-501.
Trapp, R. J., E. D. Mitchell, G.A. Tipton, D. W. Effertz, A. I. Watson, D. L. Andra, and M. A.

Magsig, 1999: Descending and non-descending tornadic vortex signatures detected by WSR-
88Ds. Wea. Forecasting, 14, 625-639.
Trapp, R. J., G. J. Stumpf, and K. L. Manross 2005: A reassessment of the percentage of tornadic
mesocyclones. Wea. Forecasting, 20, 680-687.
Wakimoto, R. M., C. Huaqing, and H. V. Murphey, 2004: The superior, Nebraska supercell during
BAMEX. Bull. Amer. Meteor. Soc., 85, 1095-1106.
Willoughby, H. E., F. D. Marks, and R. J. Feinberg, 1984: Stationary and moving convective bands
in hurricanes. J. Atmos. Sci., 22, 3189-3211.
Wilson, J. W. and D. Reum, 1986: "The hail spike": reflectivity and velocity signature. Preprints,
23rd Conf. on Radar Meteorology. Snowmass, CO, Amer. Meteor. Soc., 62-65.
Wilson, J. W. and D. Reum, 1988: The flare echo: Reflectivity and velocity signature. J. Atmos.
Oceanic Technol., 5, 197-205.
Witt, A., and S. P. Nelson, 1991: The use of single-Doppler radar for estimating maximum
hailstone size. J. Appl. Meteor., 30, 425- 431.
Witt, A., M. D. Eilts, G. J. Stumpf, J. T. Johnson, E. D. Mitchell, and K. W. Thomas, 1998: An

enhanced hail detection algorithm for the WSR-88D. Wea. Forecasting, 13, 286-303.
Zrnic, D. S., 1987: Three-body scattering produces precipitation signature of special diagnostic
value. Radio Sci., 22, 76-86.

4-133
4-134
APPENDIX A
ACRONYMS AND ABBREVIATIONS
A/D -Analog-to-Digital
AGL -Above Ground Level
AP -Anomalous Propagation
ARL -Above Radar Level
ARTCC -Air Route Traffic Control Center
ASOS -Automated Surface Observing System
ASR-9 -Airport Surveillance Radar - 9
AWIPS -Advanced Weather Interactive Processing System
BDDS -Base Data Distribution System

BIT -Built-In Test
BITE -Build-In Test Equipment
BWER -Bounded Weak Echo Region
CAPE -Conditional Available Potential Energy

CLD -Clutter Likelihood Doppler Product
CLR -Clutter Likelihood Reflectivity Product
CODE -Common Operations and Development Environment
COMET -Cooperative Program for Operational Meteorology, Education and Training
CONUS -Contiguous United States
CS -Combined Shear Product
CSU/DSU -Channel Service Unit/Data Service Unit
CR -Composite Reflectivity Product
CWA -County Warning Area
DAU -Data Acquisition Unit

dB -Decibel
DBV -Integrated Terminal Weather System Digital Base Velocity Product
dBZ -Radar Reflectivity Factor (Logarithmic Unit)
dBZe -Decibels relative to an equivalent reflectivity factor
DCZ -Deep Convergence Zone
DHR -Digital Hybrid Scan Reflectivity Product
DOC -Department of Commerce
DoD -Department of Defense
DOT -Department of Transportation
DPA -Hourly Digital Precipitation One-Hour Array Product
DQA -Data Quality Algorithm
DR -Base Reflectivity Data Array Product
DSP -Storm Total Rainfall Accumulation Product (Digital Storm Product)
DV -Base Velocity Data Array Product
DVL -High Resolution Digital Vertically Integrated Liquid Product

A-1
EET -High Resolution Enhanced Echo Tops Product
EPRE -Enhanced Precipitation Processing
ET -Echo Tops Product
ETC -Extra-tropical Cyclone
EVTS -Elevated Tornadic Vortex Signature
EWT -Environmental Wind Table
FAA -Federal Aviation Administration

FMH -Federal Meteorological Handbook
GUI -Graphical User Interface
HCI -Human Computer Interface

HDA -Hail Detection Algorithm
HI -Hail Index Product
HP -High Precipitation (Supercells)
HSP -Hardwired Signal Processor
HSR -Hybrid Scan Reflectivity Product
I -Analog In Phase
ICD -Interface Control Document
I/O -Input/output
ITWS -Integrated Terminal Weather System
ITWSDBV -Integrated Terminal Weather System Digital Base Velocity
LAN -Local Area Network

LCL -Lifting Condensation Level
LEWP -Line Echo Wave Pattern
LLWAS -Low Level Windshear Alert System
LOG -Logarithmic
LP -Low Precipitation (Supercell)
LRM -Layer Composite Reflectivity (Maximum) Product
M -Mesocyclone Product
MARC -Mid Altitude Radial Convergence
MCC -Mesocyclone Convective Complex
MCS -Mesocyclone Convective System
MCV -Mesoscale Convective Vortex
MD -Mesocyclone Detection Product
MDA -Mesocyclone Detection Algorithm
MESO -Mesocyclone
MEHS -Maximum Expected Hail Size
MHz -MegaHertz
MLOS -Microwave Line of Sight
MMI -Man Machine Interface
MPDA -Multiple PRF Dealiasing Algorithm

A-2
MPE -Multisensor Precipitation Estimator
MRU -Mesocyclone Rapid Update Product
MSCF -Master System Control Function
NCDC -National Climatic Data Center

NCEP -National Center for Environmental Prediction
ND -No Data
NEXRAD -Next Generation Weather Radar
NOAA -National Oceanic and Atmospheric Administration
NSSL -National Severe Storms Laboratory
NWS -NOAA’s National Weather Service
OFCM -Office of the Federal Coordinator for Meteorological Services and Supporting
Research
OHP -One-Hour Rainfall Accumulation Product
OPUP -Open System Principal User Processor
OSD -One-Hour Snow Depth Accumulation Product
OSI -Open Systems Interconnection
OSW -One-Hour Snow Water Equivalent Accumulation Product
OTR -One Time Request
PAT -Product Attributes Table

POH -Probability of Hail
POSH -Probability of Severe Hail
PRF -Pulse Repetition Frequency
PSP -Programmable Signal Processor
PUP -Principal User Processor
Q -Quadrature
R -Reflectivity Product
rf -Radio Frequency
RCS -Cross Section - Reflectivity Product
RDASC -RDA Status And Control
RDASOT -RDA System Operability Test
RDA -Radar Data Acquisition
RDASOT -RDA System Operability Test (Software)
REC -Radar Echo Classifier
RFC -River Forecast Center
RFD -Rear Flank Downdraft
RIJ -Real Inflow Jet
RIN -Rear Inflow Channel
RMS -Remote Monitoring Subsystem
ROC -Radar Operations Center
RPCCDS -Radar Product Central Collection Dissemination Service
RPG -Radar Product Generator
RPM -Revolutions Per Minute

A-3
RPS -Routine Product Set
S -Snow Water Equivalent

SAA -Snow Accumulation Algorithm
SBN -Satellite Broadcast Network
SCIT -Storm Cell Identification and Tracking Algorithm
SD -Snow Depth
SDU -Snow Depth (User Selectable)
SNR -Signal-to-Noise Ratio
SO -SuperOb Product
SPD -Supplemental Precipitation Data Product
SPS -Signal Processor Subsystem
SRM -Storm Relative Mean Radial Velocity (Map) Product
SRR -Storm Relative Mean Radial Velocity (Region) Product
SRV -Storm Relative Velocity Product
SS -Storm Structure Product
SSD -Storm Total Snow Depth Accumulation Product
SSW -Storm Total Snow Water Equivalent Accumulation Product
STI -Storm Tracking Information Product
STP -Storm Total Rainfall Accumulation Product
SW -Spectrum Width Product
SWA -Severe Weather Analysis Product
SWE -Snow Water Equivalent Product
SWP -Severe Weather Probability Product
SWR -Severe Weather Analysis – Reflectivity Product
SWS -Severe Weather Analysis – Radial Shear Product
SWU -Snow Water Equivalent (User Selectable) Product
SWV -Severe Weather Analysis - Velocity Product
SWW -Severe Weather Analysis - Spectrum Width Product
TAT -Task Attributes Table

TBSS -Three-Body Scatter Spike
TDA -Tornado Detection Algorithm
TDWR -Terminal Doppler Weather Radar
THP -Three-Hour Rainfall Accumulation Product
TOVER -Range Unfolding Overlay Threshold in RDA SPS Adaptation Data
TRU -Tornado Vortex Signature Rapid Update Product
TVS -Tornado Vortex Signature Product
UAM -User Alert Message Product

ULR -User Selectable Layer Composite Reflectivity Product
URC -Unit Radar Committee
USD -User Selectable Snow Depth Accumulation Product
USW -User Selectable Snow Water Equivalent Accumulation Product
USP -User Selectable Rainfall Accumulation Product
UTC -Universal Time (Coordinated)

A-4
V -Mean Radial Velocity Product
VAD -Velocity Azimuth Display Product
VCP -Volume Coverage Pattern
VCS -Cross Section – Velocity Product
VIL -Vertically Integrated Liquid Water Product
VWP -VAD Wind Profile Product
WAN -Wide Area Network

WARP -Weather and Radar Processor
WCCM -Wideband Communications Control Module
WDSS -Warning Decision Support System
WDTB -Warning Decision Training Branch
WER -Weak Echo Region
WG/DRMO -Working Group for Doppler Radar Meteorological Observations
WSR-57 -Weather Surveillance Radar-1957
WSR-88D -Weather Surveillance Radar-1988 Doppler
ZR, Z-R -Reflectivity – Rain Rate

ZS, Z-S -Reflectivity – Snowfall Rate

A-5
A-6
APPENDIX B
GLOSSARY
Adaptation Data: Adaptable parameter settings for WSR-88D hardware and software that change
their operating characteristics. These changes allow for system optimization based on
meteorological, climatological, and regional variations, as well as user preferences.
Adiabatic Process: A process in which a system does not interact with its surroundings by virtue
of a temperature difference between them. In an adiabatic process any change in internal energy
(for a system of fixed mass) is solely a consequence of work done. For an ideal gas and for most
atmospheric systems, compression results in warming, expansion results in cooling.
Alert: An operational condition or message created when a user-selected product value or

algorithm output is detected during an elevation or volume scan. Usually requested by a user in
order to be notified of a condition related to a meteorological event.
Aliasing: The process by which frequencies too high to be analyzed with the given sampling
interval appear at a frequency less than the Nyquist frequency.
Algorithm: A fixed step-by-step procedure, usually within system software, designed to

accomplish a given result; usually a simplified procedure for solving a complex problem; also a
full statement of a finite number of steps. Meteorological algorithms are designed to recognize
data patterns related to weather phenomena or threats.
Anomalous propagation: A propagation path of electromagnetic radiation that deviates from the
path expected due to refractive conditions in a standard atmosphere.
Antenna: (Also called “aerial”; sometimes the more general term “radiator” is used.) A conductor
or system of conductors for radiating and/or receiving radio energy. As used in radar, the
antenna is usually “directional;” that is, it has the property of radiating or receiving radio waves
in larger proportion in a given direction.
Archive Level I: The analog, time-domain output from the receiver. These data are useful for
detailed engineering studies. The data recording interface is located at the RDA.
Archive Level II: The digital base data output from the signal processor. The output also includes
status information required to properly interpret the data (e.g., information on synchronization,
calibration, date, time, antenna position, clutter and notchwidth maps, and operational mode).
Archive Level III: The base products and derived products/algorithm output produced by the
RPG. The Level III products are defined in Part A of this handbook.
Archive Level IV: The base products and derived products/algorithm output produced by the
RPG (as in Level III). Data recorded are selected by the user display system operator.

B-1
Associated Users: Those users with dedicated communications to a WSR-88D system.
Azimuth: A direction in terms of the 360° compass.
Base Data: (Also known as “Archive Level II” or “Level II.”) Those digital fields of reflectivity,
mean radial velocity, and spectrum width data in spherical coordinates provided at the finest
resolution available from the radar.
Batch Waveform: Contains both Surveillance and Doppler pulse trains alternating low and high
PRFs within each radial at each low elevation angle scanned to allow resolution of range
ambiguities. It is used where a high degree of clutter suppression is required, but where
contiguous waveforms are not needed.
Beam Filling: The measure of variation of hydrometeor density throughout the radar sampling
volume. If there is no variation in density, the beam is considered to be filled.
Beam Width: Angular width of the antenna pattern. Usually the width where the power density is
one-half that of the axis of the beam.
Bias: A systematic difference between an estimate of and the true value of the parameter.
Bin: Radar sample volume.
Book-End Vortices: (Also called “line-end vortices.”) With time, MCSs tend to develop vortex
pairs with opposite sense rotation at the ends of the convective line. Primarily created when the
system updraft tilts the easterly shear generated at the system's cold pool/updraft interface.
Boundary Layer: The layer of a fluid adjacent to a physical boundary in which the fluid motion is
affected by the boundary and has a mean velocity less than the free-stream value.
Bounded Weak Echo Region: A nearly vertical channel of weak radar echo, surrounded on the
sides and top by significantly stronger echoes. The BWER, sometimes called a vault, is related
to the strong updraft in a severe convective storm. It has long been found to be associated with a
supercell.
Bow Echo: A bow-shaped line of convective cells that is often associated with swaths of
damaging straight-line winds and small tornadoes. Key structural features include an intense
rear inflow jet impinging on the core of the bow, with bookend or line-end vortices on both sides
of the rear-inflow jet, behind the ends of the bowed convective segment. Bow echoes have been
observed with scales between 20 and 200 km, and often have lifetimes between 3 and 6 hours.
Bypass Map: In the absence of any operator-defined Clutter Suppression Regions, this map (built
by the RDA System Operability Test (RDASOT) software) specifies where to apply clutter
suppression.

B-2
Bright Band: The enhanced radar echo caused by the difference in radar reflectivity of ice and
water particles. This echo is interpreted as the delineation on a radar display between frozen and
liquid precipitation.
Cell: A compact region of relatively strong vertical air motion (at least 10 m s-1; 19 knots). In
radar, sometimes applied to individual radar echoes or radar echo cores of higher reflectivity.
Centroid: The center of mass of a storm echo or storm echo component.
Chinook: The name given to the foehn in western North America, especially on the plains to the
lee or eastern side of the Rocky Mountains in the United States and Canada.
Clear Air Mode: System scanning in order to gather data which will facilitate the detection of
precursors to precipitation development and non-meteorological echoes. This mode uses slower
scan rates than are used in the precipitation mode to provide increased sensitivity.
Clutter (or ground clutter): The pattern of radar echoes from fixed ground targets.
Clutter Filter Bypass Map: See Bypass Map.
Clutter Filter Notch Width Map: See Default Notch Width Map.
Clutter Suppression Region: An area defined by the user or by adaptation data where clutter
suppression is to be applied.
Comma Head: (Sometimes known as “rotating comma head.”) 1. The rounded portion of a
comma cloud system. This occurs to the left of the maximum wind speed axis and contains the
most rotation when viewed in motion. 2. The northern, rounded portion of a convective line of
storms called a bow echo and often associated with a mesocyclone.
Combined Attribute Table: A table affixed to the Composite Reflectivity product composed of
the outputs of other products and meteorological algorithms pertaining to severe convective
storms.
Combined Rankine Vortex: An idealized circular vortex of purely tangential flow in the core
which is in what is known as solid body rotation, where the rotational velocity increases directly
with radius. Outside the core is a flow which has zero vorticity because the shear profile exactly
cancels the curvature contribution. This is known as a potential vortex.
Complex Signal: In radar, a representation of the time-varying amplitude and phase of the
received signal as the real and imaginary parts of a time-varying complex number. These parts
are called the in-phase and quadrature components and are measured by coherent detection of
the received signal.
Conditional Instability: The state of a layer of unsaturated air when its lapse rate of temperature
is less than the dry-adiabatic lapse rate, but greater than the moist-adiabatic lapse rate.
Contiguous Waveform: Immediately adjacent waves of the same character.

B-3
Contiguous Surveillance (CS) Scan: A constant low PRF (long Rmax and low Vmax) scan
employed for the entire 360° sweep at low elevation angles to determine proper target location,
returned power, and to permit optimum clutter filtering.
Contiguous Doppler (CD) Scan: A constant high PRF (short Rmax and high Vmax) employed for
the entire 360° sweep. CD scanning is used at low elevation angles where range ambiguity
resolution is required to accurately determine "1st guess" velocity and spectrum width array
estimates. CD scanning is also used at high elevation angles where there is not range ambiguity
problem.
Contour Base Value: The lowest magnitude of a specific entity upon which contours are based.
Contour Interval: Value between contours.
Convergence: A measure of the contraction of a vector field.
Correlated Shear: An output of the mesocyclone detection algorithm indicating a 3-dimensional

shear region (i.e., vertically correlated) that is not symmetrical.
Covariance: A measure of the degree of association between two variables. In Doppler radars,
the argument (or angle) of the covariance of the complex signal is a measure of the Doppler
frequency.
Data Acquisition Unit (DAU): Monitors, within the WSR-88D, the state of all safety interlocks,
power supply voltages, and environmental conditions, and regularly reports to the RDASC using
an asynchronous serial line.
Data Level: The specific range of data values represented by a single pixel when the data are
presented in a pixel image format; the specific value a datum may assume.
Data Level Code: A code representing a specific data level; used to assign color values when such
are assigned.
Data Resolution: The resolution of the base data as produced by the signal processor, nominally 1
km (0.54 nm) x 1°(AZ) x 1°(ELEV) for the reflectivity values and 0.25 km (0.13 nm) x 1°(AZ)
x 1°(ELEV) for radial velocity and spectrum width values. Averaging and additional processing
may reduce these resolutions. A measure of the degree of association between two variables. In
Doppler radars, the argument (or angle) of the covariance of the complex signal is a measure of
the thing.
Dealiasing: Process of correcting for aliases in the velocity field.
Decibel (dB): A logarithmic expression for the ratio of two quantities. DBm is a decibel with
respect to 1 milliwatt.

B-4
Mathematically: dB = 10 Log (P1/P2)
dBa = 10 Log [(accumulation)/1 mm]
dBm = 10 Log (P/10-3)
dBR = 10 Log [(precipitation rate)/(1 mm/hr)]
dBZe = 10 Log (Ze)
Deep Convergence Zone (DCZ): A narrow and deep velocity signature characterized by strong
convergence along a nearly vertical interface extending from the radar horizon upward to
altitudes as high as 50,000 ft. Often associated with very damaging surface winds and related to
the Mid Altitude Radial Convergence.
Default Notch Width Map: In the absence of operator-defined Clutter Suppression Regions, the
WSR-88D system uses the Default Notch Width Map notch width definitions (stored RDA
adaptation data) to determine the amount of suppression to be applied to regions defined by the
Bypass Map.
Delta System Calibration (SYSCAL): This RDA status factor, often output at the user display
system, contains all the quantities of the radar equation but in terms of Ze plus a constant
converting receiver output power (the digital quantization factor, a2, in digits squared per mW) to
receiver power in dBm and the receiver gain, relating input to output power.
Dendritic crystal: A crystal, particularly a planar ice crystal, with its macroscopic form (crystal
habit) characterized by intricate branching structures of a treelike nature.
Derecho: A widespread convectively induced straight-line windstorm event. Specifically, the

term is defined as any family of downburst event clusters produced by an extratropical
mesoscale convective system. Derechos may or may not be accompanied by tornadoes.
Display Resolution: The area or two-dimensional product of the X and Y coordinates represented
by one picture element (pixel) of a raster scan display.
Divergence: A measure of the expansion or spreading out in a vector field.
Doppler radar: A radar that detects and interprets the Doppler effect in terms of the radial
velocity of a target. The signal received by radar from a moving target differs in frequency from
the transmitted frequency by an amount that is proportional to the radial component of the
velocity relative to the radar.
Doppler frequency shift: Also called Doppler effect. In general, the change in frequency of a
signal reaching a receiver when the receiver and the transmitting source are in motion relative to
one another.
Downburst: A strong downdraft that induces an outburst of damaging winds on or near the
ground.

B-5
Downdraft: Small-scale downward-moving air current in a cumulonimbus cloud.
Dry-adiabatic Lapse Rate: A process of temperature; the rate of decrease of temperature with
height of a parcel of dry air lifted by a reversible adiabatic process through an atmosphere in
hydrostatic equilibrium.
Dry-adiabatic Process: An adiabatic process in which no condensation of its water vapor occurs
and no liquid water is present.
Dryline: A low-level mesoscale boundary or transition zone hundreds of kilometers in length and
up to tens of kilometers in width separating dry air from moist air.
Echo: Energy backscattered from a target as seen on the radar display.
Echo Tops: The height of the greatest (in altitude) non-zero reflectivity value (greater than the
minimum significant reflectivity set in adaptation data, 18.5 dBZ is the default) for each 4 x 4
km (2.2 x 2.2. nm) grid box above the surface of the Earth.
Echo Training: Thunderstorm (or shower) cells following one after another over the same
location.
Elevation Scan: The process of the radar completing a full 360° rotation in azimuth for a specific
elevation angle.
Equivalent Radar Reflectivity (Ze): When all the assumptions (e.g., uniformly distributed liquid
water particles whose diameters meet the Rayleigh approximation) do not apply, the radar
reflectivity, Z, is expressed as Ze, the equivalent radar reflectivity. Typically expressed as: dBZ
= 10 Log Ze.
Estimate: A statement of the value of a quantity or function based on a finite number of samples.
Extratropical Cyclone: Sometimes called extratropical low, extratropical storm. Any cyclonic-
scale storm that is not a tropical cyclone, usually referring only to the migratory frontal cyclones
of middle and high latitudes.
Eye: In meteorology, usually the “eye of the storm” (e.g., hurricane, typhoon); that is, the roughly
circular area of comparatively light winds found at the center of a severe tropical cyclone and
surrounded by the eyewall.
Eyewall: A ring of cumulonimbus that encircles the eye of a tropical cyclone. In radar depictions,
the clouds must occupy at least 180° of arc to be called an eyewall.
Feature: A set of pattern vectors in close proximity.
Flash Flood: A flood that rises and falls quite rapidly, usually as the result of intense rainfall over
a relatively small area.

B-6
Frequency: The number of recurrences of a periodic phenomenon per unit time. Electromagnetic
energy is usually specified in Hertz, which is a unit of frequency equal to one cycle per second.
Gage Data Support System: A system external to the WSR-88D unit whose purpose is to collect
and transfer precipitation data from several rain gage sites to the WSR-88D unit.
Gravity Wave: (Also called “gravitational wave.”) A wave disturbance in which buoyancy (or
reduced gravity) acts as the restoring force on parcels displaced from hydrostatic equilibrium.
Gust Front: The boundary between the horizontally propagating cold air outflow from a
thunderstorm and the surrounding environmental air.
Hook Echo: A pendant, curve-shaped region of reflectivity caused when precipitation is drawn
into the cyclonic spiral of a mesocyclone. The hook echo is a fairly shallow feature, typically
extending only up to 3–4 km in height within a supercell storm before becoming part of a
Bounded Weak Echo Region (BWER).
Hydrometeor: Any product of condensation or deposition of atmospheric water vapor, whether

formed in the free atmosphere or at the earth's surface; also, any water particle blown by the
wind from the earth's surface.
In-phase (signal): The signal obtained by demodulating the received signal with a local oscillator
having the same phase and frequency as the transmitted signal.
Isolated Storm: An individual cell or group of cells that are identifiable and separate from other
cells in a given geographic area.
Kalman Filter: A linear system in which the mean squared error between the desired output and
the actual output is minimized when the input is a random signal generated by white noise.
Klystron: A power amplifier tube used to amplify weak microwave energy (provided by a radar
frequency exciter) to a high power level for a radar transmitter.
Lifting Condensation Level (LCL): (Also called “isentropic condensation level.”) The level at
which a parcel of moist air lifted dry-adiabatically would become saturated.
Low-topped Supercells: A convective storm that contains similar radar characteristics to those of
a supercell (e.g., mesocyclone, hook echo, WER, BWER), but is significantly smaller in height.
Lake-effect Snow: Localized, convective snow bands that occur in the lee of lakes when
relatively cold air flows over warm water. In the United States, this phenomenon is most noted
along the south and east shores of the Great Lakes during arctic cold-air outbreaks.
Lapse Rate: The decrease of an atmospheric variable with height, the variable being temperature,
unless otherwise specified.

B-7
Limited-Area Fine Mesh (LFM): A rectangular grid based on a polar stereographic projection.
The grid mesh length of the LFM, 1/4 LFM and 1/40 LFM at 60° N (standard latitude) and 105°
W (standard longitude) are 190.5, 47.625, and 4.7625 km (102.9, 25.7 and 2.6 nmi),
respectively.
Line Echo Wave Pattern (LEWP): A special configuration in a line of convective storms
configured like a wave which may indicate the presence of a low pressure area and the
possibility of damaging winds and tornadoes. In response to very strong outflow winds behind
it, a portion of the line may bulge outward forming a bow echo.
Low-Level Jet (LLJ): (Also called “low-level jet stream.”) A jet stream that is typically found in
the lower 2–3 km of the troposphere. At night, sometimes called a nocturnal jet. Examples are
the Great Plains low-level jet, extratropical cyclone low-level jet, African jet and the Somali jet.
Master System Control Function (MSCF): The computer-user interface at the WSR-88D system
RPG.
Master Request List: A list produced in the RPG at the start of each volume scan composed of all
algorithm outputs (data types or products) which must be generated in order to satisfy default
product requests, external user product requests, unsatisfied one-time product requests, and the
alerting function.
Mesoscale Convective Complex (MCC): A subset of mesoscale convective systems (MCS) that
exhibit a large, circular (as observed by satellite), long-lived, cold cloud shield.
Mesoscale Convective System (MCS): A cloud system that occurs in connection with an
ensemble of thunderstorms and produces a contiguous precipitation area on the order of 100 km
or more in horizontal scale in at least one direction. An MCS exhibits deep, moist, convective
overturning contiguous with, or embedded within, a mesoscale vertical circulation that is at least
partially driven by the convective overturning.
Mesoscale Convective Vortex: A warm core mid-altitude mesoscale vortex often produced as an
MCC or MCS decays; often associated with a recurrence of convection.
Mean Radial Velocity: The component of motion of the target toward or away from the radar.
Mesocyclone: A 3-dimensional region in a storm that contains strong cyclonic vertical vorticity
(rotates cyclonically) and is closely correlated with severe weather.
Mesoscale: On a scale of 4 km to 400 km (2.2 nm to 215 nm).
Microburst: Small downburst, 1 to 4 km (0.54 to 2.2 nm) in outflow size, with peak winds lasting
2 to 15 minutes.
Mid Altitude Radial Convergence (MARC): A deep, mid-level, convergent velocity signature
related to the Deep Convergence Zone (but confined to mid-levels) and found to be a Doppler

B-8
radar-based precursor of damaging straight-line winds in a linear MCS or bowing convective
system.
Mie Scattering: Scattering of electromagnetic waves by homogeneous spheres of arbitrary size,

named after Gustav Mie (1868–1957), whose theory of 1908 explains the process. Often used in
reference to radar waves scattered by particles with small diameters as compared to radar
wavelength.
Mini-Supercell: A convective storm that contains similar radar characteristics to those of a

supercell (e.g., mesocyclone, hook echo, WER, BWER), but is significantly smaller in height
and width.
Misoscale: On a scale of 40 m to 4 km (130 ft to 2.2 nm).
Mixing Ratio: The ratio of the mass of a variable atmospheric constituent to the mass of dry air. If
not otherwise indicated, the term normally refers to water vapor.
Moist-adiabatic Process: (Also known as “saturation-adiabatic process.”) An adiabatic process

for which the air is saturated and may contain liquid water.
Nyquist Interval: (Also known as “Nyquist velocity.”) The maximum time interval between
equally spaced samples of a signal that will enable the signal waveform to be completely
determined. Also known as the absolute value of the maximum unambiguous velocity that can
be measured by a Doppler radar (e.g., 50 kts).
Nyquist Co-Interval: The full range of the Nyquist interval (e.g., +/- 50 kts).
Operational Mode: A combination of one or more volume coverage patterns and product mixes
tailored to one or more meteorological situations.
One-Time Request (OTR): Requests of products made on a one-time basis via Associated Users.
Overhang: A storm has overhang if the edge of the storm component at a given height range
(mid-levels) extends outward beyond the edge of the storm component at the lowest elevation by
a specified distance.
Pattern Vector: A pattern vector is formed by a series of azimuthally-adjacent sample volumes of

increasing or decreasing Doppler velocity.
Pedestal: In radar, a device for supporting and positioning the antenna. Typically, the pedestal
allows the azimuth and elevation angles of the antenna to be controlled separately or in a
coordinated way to permit different methods of scanning.
Point Clutter Rejection: The rejection or removal of echoes having the characteristics of point
targets.
Polarization: With respect to a transverse electromagnetic wave, the correlation between two
orthogonal components of its electric (or, equivalently, magnetic) field.

B-9
Polar Coordinates: A system of coordinates in which a point is isolated by its distance (range)
and angular direction (azimuth) from a fixed reference point. For radars, this reference point is
usually the location of the antenna.
Precipitation Mode: System scanning at variable rates to accommodate the greatest number of
elevation angles, thus sampling the full radar volume in order to detect precipitation in all its
forms.
Precipitable Water: (Also called “precipitable water vapor.”) The total atmospheric water vapor
contained in a vertical column of cross-sectional area extending between any two specified
levels, commonly expressed in terms of the height to which that water substance would stand if
completely condensed and collected in a vessel of the same unit cross section.
Product: Output of the WSR-88D receiver in the form of the three base moments (reflectivity,
mean radial velocity, and spectrum width) formatted by the RPG as base, derived, or algorithm-
processed alphanumeric messages, graphic images, or graphic overlays for presentation on a
user display system.
Product Distribution Function: This RPG function is the distribution of all products requested
via RPS Lists and One-Time Requests.
Product Generation Function: (Sometimes called the “Product Generation and Control List.”)
Each volume scan the RPG combines the products requested via the RPS list, One-Time
requests, alert generated products, and the default list for the weather mode into one list called
the Master Request List. This list is what the system uses to determine what products are
generated.
Product Generation Table: A table created in the RPG each volume scan. The table is
composed of products that are to be generated independent of user requests at the RPG.
Separate tables for Precipitation, Clear Air, and Maintenance modes exist.
Product Storage Load Shedding: Refers to the removal of a product from the RPG product
database prior to the product's expiration time. This is required when the number of products in
the RPG database exceeds a specified number (~7000 products). In order to experience load
shedding, many products need to be generated with long storage times (the maximum storage
time is 360 minutes).
Product Resolution: The smallest spatial increment of data element that is distinguishable in a
product.
Propagation: Transmission of electromagnetic energy as waves through or along a medium.
Pulse: A single short duration transmission of electromagnetic energy.
Pulse Severe Storm: A convective storm characterized by a single, strong, updraft pulse
producing a short-lived period of large hail or damaging winds at the surface.

B-10
Pulse Width: The linear distance in range occupied by an individual broadcast from a radar.
Quadrature (signal): In radar systems, an orthogonal relationship between two coherent signals
in which the phase of one signal is offset by 90° from the phase of the other. Two signals in
quadrature may be regarded as a single complex signal. In Doppler radar, the signal is
composed of the in-phase and quadrature components.
Radar: (Coined word for “radio detection and ranging.”) An electronic instrument used for the
detection and ranging of distant objects of such composition that they scatter or reflect
microwave radio energy. A radar consists of a transmitter, receiver, antenna, display, and
associated equipment for control and signal processing.
Radar Data Acquisition System Operability Test (RDASOT): An off-line program designed to
determine the status of the hardware and to enhance its maintainability. RDASOT, executed
from the system console, consists of diagnostic tests, calibration tests, and maintainability aids.
Radar horizon: The locus of points at which direct rays from a radar transmitter/antenna become
tangential to the earth's surface. The radar horizon extends beyond the geometric and visible
horizons in conditions of normal atmospheric refraction. It may be decreased or increased in
particular cases as standard propagation is replaced by substandard or superstandard propagation
respectively. Beyond the radar horizon, surface targets cannot be detected under rf atmospheric
conditions although significant radar power is sometimes detected in the diffraction zone below
the horizon.
Radome: A dome used to cover the antenna assembly of a radar to protect it from the effects of
weather.
Range Aliasing: (Also called range folding.) In radar meteorology, a sampling problem that
arises when echoes located beyond the maximum unambiguous range (Rmax) are received as if
they were within this radar range. A radar ordinarily computes range to targets by measuring the
time interval between the transmission of a pulse and the receipt of the returned signal, assuming
that the signal was associated with the pulse just transmitted. However, depending on the pulse,
the returned signal may be associated with one of several pulses transmitted prior to the latest
one. Therefore, a returned signal, indicated as originating at range r, could have originated at r
+ Rmax (second-trip echo), or r + 2Rmax (third-trip echo), etc.
Range Dealiasing: (Also known as “range unfolding.”) The process or processes of removing
range ambiguity in apparent range of a multitrip target; that is, to assign the correct target range.
Rayleigh Scattering: Approximate theory for electromagnetic scattering by small particles named
for Lord Rayleigh (John William Strutt, 1842–1919). Commonly used in radar theory referring
to small particles as compared to radar wavelength.
Rear Flank Downdraft: A downdraft almost exclusively associated with supercell storms found
along the rear portion (facing in the direction of storm motion) of the storm and associated with
the mesocyclone and often, tornadoes. Sometimes responsible for damaging surface winds.

B-11
Rear Inflow Jet (RIJ): A mesoscale circulation feature in which a system-relative current of air
enters and flows through the stratiform precipitation region of MCS from the rear. The rear
inflow jet forms in response to the upshear-tilting of the convective circulation, as the horizontal
buoyancy gradients along the back edge of the system create a circulation that draws mid-level
air in from the rear. The rear inflow jet supplies potentially cold and dry mid-level air that aids
in the production of convective and system-scale downdrafts.
Rear Inflow Notch (RIN): A channel of weak echo extending from the rear into a convective
storm line. Often associated with the rear inflow jet.
Receiver: An instrument used to detect the presence of, and to determine the information carried
by, electromagnetic radiation. A receiver includes circuits designed to detect, amplify, rectify,
and shape the incoming radio-frequency signals received at the antenna.
Reflectivity: A measure of the fraction of radiation reflected by a given surface; defined as a ratio
of the radiant energy reflected to the total that is incident upon that surface.
Refraction: Changes in the direction of energy propagation (due to changes in speed) as a result
of density changes within the propagating medium.
Refractive Index: A measure of the amount of refraction. Numerically equal to the ratio of wave
velocity in a vacuum to wave speed in the medium.
Routine Product Scheduler: Within the RPG, this function publishes the Master Request List for
all algorithms to use in processing and generation of each algorithm’s output. See Master
Request List.
Routine Product Set (RPS): A product list requested on a routine basis from user display systems
and external users.
Scatterer: Any object capable of reflecting the radar signal.
Sectorized Hybrid Scan: A single reflectivity scan composed of data from the lowest four
elevation scans. Close to the radar, higher tilts are used to reduce clutter. At further ranges,
either the maximum values from the lowest two scans are used or the second scan values are
used alone.
Segment: As applied in the Storm Segments algorithm, segments are defined as runs of contiguous
radar sample volumes greater than or equal to the minimum reflectivity threshold and having a
combined length greater than or equal to the segment length threshold.
Severe Storm: A storm with a tornado, surface hail >3/4 inch, or wind gusts >50 knots; the storm
could contain all three.
Shear: The rate of change of the vector wind in a specified direction normal to the wind direction.
Vertical shear is the variation of the horizontal wind in the vertical direction.

B-12
Sidelobe: Secondary radiated energy maximum other than the radar main beam. Typically
contains a small percentage of energy compared to the mainlobe.
Signal Processor: A computer processor used to apply a series of algorithms to the output of the
receiver in order to estimate the spectral moments contained in the received backscattered signal.
Signal to Noise Ratio: A ratio that measures the comprehensibility of data, usually expressed as
the signal power divided by the noise power.
Spectrum Width: A measure of dispersion of velocities within the radar sample volume.
Standard deviation of the mean radial velocity spectrum.
Split Cut (CS/CD) Scan: While staying at a particular elevation angle, there is one full rotation
using the CS waveform, followed by one full rotation using the CD waveform.
Stratiform: Descriptive of clouds or precipitation of extensive horizontal development, as

contrasted to the vertically developed convective clouds or precipitation types.
Storm: Any disturbed state of the atmosphere, especially as affecting the Earth’s surface, and
strongly implying destructive and otherwise unpleasant weather. Storms range in scale from
tornadoes and thunderstorms, through tropical cyclones, to widespread extra-tropical cyclones.
Supercell: An often dangerous convective storm that contains radar characteristics such as the
hook echo, WER, and BWER but also contains a deep, persistent mesocyclone characterized
most often by cyclonic vorticity and closely associated with the dominant storm updraft and
RFD. Variations include the “Low Precipitation” (LP), Classic, and “High Precipitation” (HP)
supercells. These storms are often long-lived, often move somewhat differently than other non-
severe storms in their environment, and commonly produce severe weather.
Subrefraction: The propagation of radar energy under conditions of substandard refraction in the
atmosphere; that is, refraction by an atmosphere or section of the atmosphere in which the
refractive Index decreases with height. Substandard propagation produces less than normal
downward bending, or even upward bending, of the radar beam as it travels through the
atmosphere.
Superrefraction: The propagation of radar energy under conditions of superstandard refraction

(superrefraction) in the atmosphere; that is, refraction by an atmosphere or section of the
atmosphere in which the refractive index decreases with height and range. Superstandard
propagation produces greater than normal downward bending of radar beams as they travel
through the atmosphere, giving extended radar horizons and increased radar coverage. It is
caused primarily by propagation through layers near the earth's surface in which the dewpoint
temperature is rapidly decreasing or the temperature increases with height.
Surveillance Waveform: A constant low PRF (long Rmax and low Vmax) scan employed for the
entire 360° sweep to determine proper target location and returned power. This PRF is
unambiguous over the observing domain (range). Generally used as part of a split cut consisting
of a surveillance waveform followed immediately by a range-ambiguous Doppler scan. (See
also Contiguous Surveillance (CS) Scan).

B-13
Target: Precipitation or other phenomena that produce echoes.
Three-body Scatter Spike: (Also called a “flare echo.”) A long, narrow, weak reflectivity echo
artifact sometimes found extending down radial from highly reflective echo cores. Indicative of
large hail and caused by forward Mie scattering or radar signals, reflecting from the hail core, to
the ground, back to the hail core, and back to the radar.
Tilt: A storm is said to have tilt if a line connecting the centroid of a mid-level storm component
to the centroid of the lowest storm component is to the right or rear of the direction of movement
of the storm.
Tornadic Vortex Signature (TVS): The Doppler velocity signature of a tornado or incipient
tornado-like circulation within any scanned elevation angle. As the signature occurs when the
radar beam is wider than the vortex, the measured Doppler velocities are weaker than the
rotational velocities within the vortex and the apparent core diameter is larger than that of the
vortex. The signature, which may extend throughout a considerable vertical depth, is ideally
characterized by extreme Doppler velocity values of opposite sign separated in azimuth by the
equivalent of one beamwidth.
TOVER: An adaptable parameter (power difference) used in the radar range dealiasing algorithm.
The parameter is applied in comparing power returned from a range gate and those separated by
the unambiguous range or multiples of that range from the range gate in question.
Transmitter: A device used for the generation of signals of any type and form that are to be
transmitted. In radio and radar, it is that portion of the equipment that includes electronic
circuits designed to generate, amplify, and shape the radio frequency energy that is delivered to
the antenna where it is radiated out into space.
Tropical Cyclone: The general term for a cyclone that originates over the tropical oceans. This
term encompasses tropical depressions, tropical storms, hurricanes, and typhoons.
Turbulence: Random and continuously changing air motions that are superposed on the mean
motion of the air.
Unambiguous Range: The range to which a transmitted pulse wave can travel and return to the
radar before the next pulse is transmitted.
Uncorrelated Shear: An output of the Mesocyclone Detection algorithm indicating a region of

shear that is large and symmetrical but not vertically correlated.
Unit Radar Committee: A committee formed at WSR-88D sites where there is more than one
Associated Principal User.
Updraft: A small- scale current of air with marked vertical motion that is upward moving.

B-14
Velocity Aliasing: (Also called velocity folding.) A basic sampling problem arises when the
unambiguous velocity sampling interval is less than the full range of naturally occurring
velocities, causing the erroneous appearance of higher velocities within the sampling interval.
This phenomenon occurs in Doppler velocity measurements when the maximum unambiguous
velocity interval (±Vmax) is less than the full range of velocities being measured. Any true
velocity, V, appears within the interval from -Vmax to +Vmax, with the value V , which is related to
the true velocity by V = V ± 2nVmax, where n is an integer.
Velocity Dealiasing: (Also called “velocity unfolding.”) A process or processes by which

ambiguous velocities are assigned their correct unambiguous value.
Vertically Integrated Liquid (VIL): Vertical integral of liquid water content obtained from radar
observations at different elevation angles within a precipitation volume; has dimensions of mass
per unit area. The liquid water content is computed from the equivalent radar reflectivity factor,
Ze, using the Marshall-Palmer drop-size distribution.
Volume Coverage Pattern: A volumetric sampling procedure designed from the surveillance of
one or more particular meteorological phenomena.
Volume Scan: The process of completing a series of specified scans in a specific sequence.
Vortex: In its most general use, any flow possessing vorticity. More often the term refers to a
flow with closed streamlines.
Warning: A message or condition created when an adverse situation is detected by the system in
the WSR-88D hardware or software.
Waveform: The pictorial representation of the shape of a wave showing the amplitude variations
as a function of time. Often used to represent other wave properties, as well.
Waveguide: A type of conductor used to carry VHF or microwave energy from one point to
another. Most waveguides are hollow (rectangular or circular) and carefully dimensioned
according to the frequency and energy to be conducted. In radar, it provides a path for the
microwave energy between the antenna and the transmitting and receiving systems.
Weak Echo Region (WER): Within a convective echo a localized region of weak radar echo that
is bounded on one side and above by strong echo and associated with the strong updraft region.
It is located on the low-altitude inflow or updraft side of the storm.
Wind Shear: The local variation of the wind vector or any of its components in a given direction.
WSR-88D System: The summation of all hardware, software, facilities, communications,

logistics, staffing, training, operations, and procedures specifically associated with the
collection, processing, analysis, dissemination, and application of data from the WSR-88D unit.
WSR-88D Unit: The combination of one RDA, one RPG, and all associated OPUPs and
interconnecting communications.

B-15
Ze: See Equivalent Radar Reflectivity.

B-16

Dopp RDR WX Obs 06

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Dopp RDR WX Obs 06

Încărcat de

Drepturi de autor:

Formate disponibile

U.S.

DEPARTMENT OF COMMERCE/ National Oceanic and Atmospheric Administration

OFFICE OF THE FEDERAL COORDINATOR FOR

FEDERAL METEOROLOGICAL HANDBOOK NO. 11

VADM CONRAD C. LAUTENBACHER, JR., USN (RET.) MR. RANDOLPH LYON

DR. SHARON HAYS (Acting) MR. CHARLES E. KEEGAN

DR. RAYMOND MOTHA MR. DAVID MAURSTAD (Acting)

MR. JAMES B. HARRISON, Executive Secretary

THE INTERDEPARTMENTAL COMMITTEE FOR

MR. SAMUEL P. WILLIAMSON, Chairman MS. LISA BEE

MS. ANDREA PETRO

MR. JAMES B. HARRISON, Executive Secretary

FEBRUARY 2006 FMH-11-PART D

The original Federal Meteorological Handbook, Number 11 (FMH-11), DOPPLER RADAR

Part A - System Concepts, Responsibilities, and Procedures (December 2005)

Part B - Doppler Radar Theory and Meteorology (December 2005)

Part C - WSR-88D Products and Algorithms (Expected in 2006)

Part D - WSR-88D Unit Description and Operational Applications (February

FEBRUARY 2006 FMH-11-PART D

FEBRUARY 2006 FMH-11-PART D

CHAPTER 2. UNIT DESCRIPTION

2.2.1 Antenna 2-1

2.3 Radar Product Generator 2-8

2.3.1 Master System Control Function 2-8

2.4 Associated Principal User Display Systems 2-13

2.4.1 Advanced Weather Interactive Processing System 2-13

FEBRUARY 2006 FMH-11-PART-D

2.5 Communications 2-14

2.5.1 Wideband Communications 2-14

CHAPTER 3. UNIT FUNCTIONAL FLOW

3.2.1 Handle Maintenance Console 3-1

3.3 Radar Product Generation 3-8

3.3.1 System Control 3-10

CHAPTER 4. OPERATIONAL APPLICATIONS

4.2.1 Ground Clutter 4-1

4.3 Non-Precipitation Radar Echoes 4-13

4.3.1 Recognition of Non-Precipitation Radar Echoes 4-15

4.4 Boundaries 4-15

4.4.1 Recognition of Boundaries 4-15

FEBRUARY 2006 FMH-11-PART-D

4.5 Cloud Layers 4-19

4.5.1 Recognition of Cloud Layers 4-20

4.6 Pre-Convective Development 4-22

4.6.1 Recognition of Pre-Convective Development 4-22

4.7 Convective Storms 4-22

4.7.1 Single Cell Storms 4-22

4.8 Hail 4-65

4.8.1 Recognition of Hail Potential 4-65

4.9 Extratropical Cyclones 4-70

4.9.1 Evolution of Stratiform Clouds and Precipitation 4-71

4.10 Winter Storms 4-75

4.10.1 Onset of Surface Precipitation 4-75

4.11 Tropical Cyclones 4-89

4.11.1 Embedded Convective Storms and Tornadoes 4-91

FEBRUARY 2006 FMH-11-PART-D

4.12 Flash Floods 4-99

4.12.1 Recognition of a Flash Flood Event 4-99

4.13 Non-Convective High Wind Events 4-106

4.13.1 Recognition of High Wind Events 4-106

4.14 Low-Level Jets 4-107

4.14.1 Recognition of Low-Level Jets 4-107

4.15 Wind Shear 4-110

4.15.1 Recognition of Wind Shear 4-110